What we build next.
Suggest ideas, vote for the ones you need and follow their status. The most voted move up the list.
Agents: Crawl a whole site or sitemap with include and exclude patterns
Chatbase crawls a website or sitemap and auto-resyncs weekly (https://www.chatbase.co/docs/user-guides/chatbot/data-sources.md); Voiceflow manages sitemaps with include/exclude patterns and a refresh schedule (https://www.voiceflow.com/docs/documentation/build/importing-data-sources.md). We allow 20 single URLs (AGENT_MAX_URLS) re-synced daily. A team moving from an old help center on another domain has hundreds of pages. First scope: add a site or sitemap URL, discover pages up to a cap (e.g. 500) with include/exclude path patterns, fetch through safeFetch with a crawl budget per cron tick, show the page list with status, and re-sync weekly.
Track: Create issues from email with a per-project intake address
Linear gives each team a unique intake email; new emails become issues in triage with sender and attachments, and replies do not create new issues (https://linear.app/integrations/create-issues-via-email, https://linear.app/docs/customer-requests). Track already has a triage queue for non-app sources (lib/track-bulk.ts, source column) but no email entry point. Customers and non-users report bugs by email, so this turns Track into a lightweight help desk for small teams. First scope: a per-project address using the existing inbound email handling, issues created with source email and triage on, sender stored on the issue, attachments copied. Priority: low. Effort: M. Compared with Linear.
- 1
Track: Share saved views with the team and allow workspace-level views
Linear views can be personal or shared with a team or the workspace, and filtered views can be shared by URL or shortcut (https://linear.app/changelog, Jul 2, 2026 entry). Track saved views are filtered by user_id in listIssueViews (lib/track-data.ts), so each teammate has to rebuild the same Bugs this week view, and views only exist inside one project. First scope: a visibility flag (personal or team) on track_issue_views, shared views listed for all project members, and saved views on the cross-project Issues page. Priority: medium. Effort: S. Compared with Linear.
Notes: Add a personal inbound email address that turns forwarded email into notes
Evernote gives every user a private address; forwarded emails become notes, and the subject line accepts @notebook and #tag shortcuts, up to 200 emails per day (https://help.evernote.com/hc/articles/209005347). We have no inbound capture besides the bookmarklet (app/api/notes has no email route), although the suite already sends and receives email through Resend and CRM email sync. Forwarding receipts, confirmations and threads is one of the fastest capture habits and works from any phone mail client. First scope: a per-person secret address shown in Notes settings, an inbound webhook that creates a note with subject as title and body as blocks, and @notebook and #tag parsing in the subject. Priority: medium. Effort: M. Compared with Evernote.
bug en new ideas
dejé varias ideas y no se ven en el board de new
CRM: Two-way sync with HubSpot so a team can use both during a migration
HubSpot: Data sync keeps HubSpot and another app in step in one or both directions, with field mappings, filters and conflict rules (https://knowledge.hubspot.com/integrations/connect-and-use-hubspot-data-sync). Today: lib/crm-import-jobs.ts:16-24 (re-import updates through crm_external_ids); lib/groups-hubspot-sync.ts:1-25 syncs company properties for Lines only. Moving a sales team takes weeks; during that time both CRMs must show the same contacts, companies, deals and activity. First scope: connect a HubSpot private app token from CRM settings, preview what would sync, keep contacts, companies and deals fresh from HubSpot, then write changes back with per-property conflict rules, a status page with errors and retry, and a switch to pause everything.
CRM: Start new workspaces with the HubSpot default pipeline and a HubSpot-style setup option
HubSpot: The default Sales pipeline has seven stages with preset probabilities (https://knowledge.hubspot.com/object-settings/set-up-and-customize-pipelines). Today: lib/seed.ts:20-22, :34-40. A team that arrives from HubSpot expects the stages and names it knows. First scope: seed the seven HubSpot default stages with their probabilities for new workspaces (or as a Use HubSpot defaults button in pipeline settings), and review menu names against HubSpot (goals, recycle bin, properties). Also covers the parity rows nav-structure.
CRM API: search with filters, batch endpoints, and endpoints for properties and owners
HubSpot: CRM API: objects, properties, associations, owners, pipelines, lists, search and batch endpoints (https://developers.hubspot.com/docs/api-reference/latest/crm/using-object-apis.md). Today: app/api/v1/crm/{contacts,companies,deals,leads,objects,tasks,activities,products,quotes,pipelines,associations,tags,sequences,forms,email-templates}. Integrations written for HubSpot rely on search and batch calls; without them every sync is one request per record. First scope: POST search on contacts, companies and deals with the filter rules of the UI, batch create, update and read of up to 100 records, and read endpoints for property definitions and workspace members.
CRM: Audit log of exports, imports, bulk deletes and settings changes
HubSpot: Account audit logs list who exported, deleted, imported or changed settings, with an export history page (https://knowledge.hubspot.com/account-management/view-and-export-account-activity-history). Today: lib/crm.ts:433 property events; lib/crm-trash.ts; no CRM audit module among lib/*-audit.ts (board, chat, docs, groups, track). When data leaves or disappears an admin needs to know who did it. First scope: a CRM audit table and page (admins only) with exports, imports, bulk edits and deletes, merges, permission and pipeline changes, with who and when, never the record contents.
CRM: Permissions per object for view, edit and delete (all, team, owned), with export and import switches
HubSpot: Permissions per object: view, edit and delete, each set to everything, team only, owned only or none, plus export, import and bulk delete, saved as permission sets (https://knowledge.hubspot.com/user-management/hubspot-user-permissions-guide). Today: lib/crm-visibility.ts:19 one MEMBER_VISIBILITIES setting for the workspace; lib/roles.ts:11. A HubSpot admin limits who can edit or delete other people records and who can export the database. Today whoever sees a record can delete it. First scope: per role or per person, for contacts, companies, deals and tasks: view, edit and delete each set to all, team, owned or none, plus switches for export, import and bulk delete, enforced in the same place as ownerScope.
CRM: More workflow actions: rotate owner, create a deal, remove from a list, several branches
HubSpot: Actions include Rotate record to owner, Create record, Edit record (incl. associated records), Add to and Remove from static list, Send internal notifications, branches with many paths (https://knowledge.hubspot.com/workflows/choose-your-workflow-actions). Today: lib/crm-workflows.ts:57-69 ACTION_TYPES; branch semantics at :917-925. These are the actions HubSpot sales workflows use most after tasks and emails. First scope: rotate owner among a team (reusing round-robin assignment), create a deal or a task on the associated record, remove from a static list, set a property on associated records, and a branch with more than two paths.
CRM: Enroll in a workflow by filter criteria, with re-enrollment, for companies too
HubSpot: Enrolment triggers can be filter criteria (property values, list membership, activity) evaluated continuously, with optional re-enrolment and enrolling existing records (https://knowledge.hubspot.com/workflows/set-your-workflow-enrollment-triggers). Today: Triggers are events with a small config (form, score threshold) (lib/crm-workflows.ts:36-54, :816-825); no criteria or re-enrolment setting. HubSpot workflows start with When a record meets these conditions. First scope: trigger conditions built with the list filter builder on contact, company and deal workflows, checked on record changes and on a schedule, an Enroll existing records option, and a re-enrollment switch. Also covers the parity rows workflow-company-objects.
CRM: Delays in workflows
HubSpot: Delay actions wait a set time, until a date, a day and time, or until an event happens (https://knowledge.hubspot.com/workflows/choose-your-workflow-actions). Today: No delay or wait in ACTION_TYPES (lib/crm-workflows.ts:57-69); only sequences have wait steps (lib/crm-sequences.ts:23). Most HubSpot workflows wait between steps (send, wait 3 days, create a task). Without delays they cannot be rebuilt. First scope: a delay action (minutes, hours, days, until a date property) backed by a scheduled run claimed from the minute cron, visible in the run log, cancelled when the record no longer exists.
CRM: Two-way Google Calendar sync and scheduling a meeting from the record
HubSpot: With Google Calendar or Outlook Calendar connected, meetings created on a record go to the calendar with invites, and calendar events with contacts are logged on the record (https://knowledge.hubspot.com/integrations/use-hubspots-integration-with-google-calendar-or-outlook-calendar). Today: lib/crm-integrations.ts:166 (Cal bookings), lib/crm-outlook-sync.ts; no Google Calendar read in lib/crm-email-sync.ts. Meetings are the main logged activity for most sales teams and today only Cal bookings and Outlook events arrive. First scope: read Google Calendar events with contacts into the timeline (the Gmail connection already exists), and a Meeting action on the record that creates the event with attendees on the connected calendar.
CRM: BCC address to log email from any client, and a never-log list
HubSpot: Each account has a BCC address (logs outgoing mail) and a forwarding address (logs received mail) that work from any mail client (https://knowledge.hubspot.com/connected-email/log-email-in-your-crm-with-the-bcc-or-forwarding-address). Today: No BCC or forwarding address in lib/crm-email.ts, lib/crm-email-sync.ts or lib/crm-extension.ts. Not everyone connects an inbox; the BCC address is the fallback HubSpot users know, and the never-log list keeps private and internal mail out. First scope: a per-workspace inbound address that logs the message on the contacts in To and Cc, plus a list of addresses and domains that sync, the extension and BCC all skip. Also covers the parity rows email-never-log.
CRM: Named task queues, a tasks table with views, and reminders
HubSpot: Task queues: named queues a rep fills and then works one task after another; any filtered task view can be started (https://knowledge.hubspot.com/tasks/create-tasks). Today: lib/crm-work-queue.ts:11-12 scopes due, overdue, today; components/crm/QueueFilters.tsx (scope and owner). Reps organise their day in HubSpot with queues (calls to make, follow-ups) and the tasks table. First scope: queues a person creates and adds tasks to, Start tasks from any filtered task view, a table view of tasks with type, priority, queue and associated record columns and bulk edit, a reminder time on a task and a daily email with the tasks of the day. Also covers the parity rows tasks-index, task-reminders.
CRM: Tickets object with a support pipeline
HubSpot: Tickets are a standard CRM object with a support pipeline (New, Waiting on contact, Waiting on us, Closed), a board and associations (https://knowledge.hubspot.com/object-settings/set-up-and-customize-pipelines). Today: No ticket table among the CRM tables (lib/crm-schema.ts, lib/schema/baseline.ts:69-380). Tickets exist only inside Lines (app/api/v1/groups/tickets). Teams that use HubSpot tickets for onboarding or support cannot leave without them. First scope: tickets with name, pipeline and status, priority, owner, source and description, a board and a table, associations with contacts, companies and deals, the timeline, and the HubSpot default support pipeline (New, Waiting on contact, Waiting on us, Closed).
CRM: Associate an activity with several contacts, companies and deals
HubSpot: An activity can be associated with many contacts, companies, deals and tickets; logging on a contact also associates its primary company and up to five recent open deals (https://knowledge.hubspot.com/connected-email/log-email-in-your-crm-with-the-bcc-or-forwarding-address). Today: crm_activities has single contact_id, company_id and deal_id columns (lib/schema/baseline.ts:172-174). A meeting with three people from one account shows on one contact today. HubSpot shows it on all of them and on the deal, and imported activities lose links. First scope: an association table for activities (keeping the current columns as the primary link), auto-association with the contact company and open deals when logging, and an Associated with picker in the composer.
CRM: Admin-defined association labels, also between deals and contacts
HubSpot: Association labels describe the relationship between any two records (Decision maker, Billing contact...); admins create single and paired labels per object pair (https://knowledge.hubspot.com/object-settings/create-and-use-association-labels). Today: lib/crm-constants.ts:38-45 six fixed contact to company labels; lib/crm-assoc.ts:41-44 rejects others; crm_deal_contacts has no label (lib/schema/baseline.ts:153-157). Deal contact roles (decision maker, champion, billing) are how HubSpot teams run multi-threaded deals, and the sync has to keep them. First scope: labels defined per object pair in settings (the six current ones become the defaults for contact and company), a label on deal to contact and deal to company links, shown and editable on the association cards.
CRM: One properties page per object with built-in and custom properties, search and descriptions
HubSpot: Settings › Properties lists every default and custom property of an object with search, filters by group, type and creator, usage, archive and restore, and export (https://knowledge.hubspot.com/properties/create-and-edit-properties). Today: components/crm/CustomFieldsPanel.tsx; lib/crm.ts:3119-3200. Read in code, not seen running. Admins coming from HubSpot manage everything from Settings › Properties. First scope: list built-in and custom properties together per object with type, group, description and where they are used, with search; descriptions show as help text next to the field.
CRM: Let admins edit the options of lead status, source, priority, industry and deal type
HubSpot: Lead status is an editable dropdown with eight defaults: New, Open, In progress, Open deal, Unqualified, Attempted to contact, Connected, Bad timing (https://knowledge.hubspot.com/records/use-lifecycle-stages). Today: lib/crm-constants.ts:14-22 LEAD_STATUSES is a const list. Every HubSpot portal has its own lead status list; a fixed list forces the team to change process during the migration and makes the sync lose values. First scope: store the options of built-in dropdowns per workspace (seeded with the HubSpot defaults, Attempted to contact included), editable in Settings › Data, with a migration of existing values. Also covers the parity rows standard-property-options.
CRM: Editable lifecycle stages that move on their own when a deal is created or won
HubSpot: Eight default lifecycle stages (Subscriber, Lead, Marketing qualified lead, Sales qualified lead, Opportunity, Customer, Evangelist, Other); admins can add and edit stages (https://knowledge.hubspot.com/records/use-lifecycle-stages). Today: lib/crm-constants.ts:5-13 LIFECYCLE_STAGES is a const list. HubSpot moves the contact and company to Opportunity and Customer automatically; without it lifecycle reports go stale after the migration. First scope: add the Other stage, let admins add and rename stages, and a setting (on by default) that sets Opportunity on deal create and Customer on deal won, never moving a record backwards. Also covers the parity rows lifecycle-auto-sync.
CRM: Add the HubSpot default properties that are missing on contacts, companies and deals
HubSpot: Contacts ship with about 200 default properties, including Mobile phone number, Street address, State/Region, Postal code, Time zone, LinkedIn URL, Number of times contacted and the Became a ... date for each lifecycle stage (https://knowledge.hubspot.com/properties/hubspots-default-contact-properties). Today: lib/schema/baseline.ts:69-87, lib/crm.ts:756-787 ContactInput, lib/crm-activity-dates.ts:1-12. Imports and the sync have to map these somewhere, and reps look for them. First scope: contacts get mobile phone, state or region, street address, postal code, time zone and LinkedIn URL; companies get annual revenue, type, state or region, postal code, street address and time zone; deals get deal type, next step and closed won reason. All show in forms, columns, filters, import mapping and the API. Also covers the parity rows default-company-properties, default-deal-properties.
CRM: Create the company from the contact email domain when it does not exist
HubSpot: With the setting on, HubSpot creates a company from the contact email domain and associates it automatically. Today: lib/crm.ts:816-837 findCompanyByDomain links only; no create. HubSpot teams rely on companies appearing on their own. First scope: a workspace setting (off by default) that creates the company from a business domain on contact create, using the existing enrichment to name it.
CRM: Additional email addresses on a contact
HubSpot: A contact can hold a primary email plus additional emails; all of them match incoming mail, forms and dedupe (https://knowledge.hubspot.com/records/add-multiple-email-addresses-to-a-contact). Today: One email column per contact (lib/schema/baseline.ts:74). People change jobs and write from two addresses; HubSpot keeps them on one contact and matches mail and dedupe on all of them. First scope: a secondary emails list on the contact, used by email sync matching, dedupe, merge (the losing email becomes a secondary one) and search.
CRM: Stop a second contact with the same email from being created
HubSpot: HubSpot deduplicates contacts by email: creating a contact with an email that exists is blocked in the form and updates the existing record in imports, forms and the API (https://knowledge.hubspot.com/records/deduplication-of-records). Today: createContact inserts without checking the email (lib/crm.ts:827-861); the index on email is not unique (lib/schema/baseline.ts:90). Only the CSV import and the API migration dedupe (lib/crm-import.ts:439). Email is the identity of a contact in HubSpot, and the two-way sync needs the same rule on our side or every sync creates duplicates. First scope: the create form and quick create show the existing contact and block the save; the API returns 409 with the existing id; forms, Cal bookings and chats update the existing contact; a unique index per workspace on the lowercased email once existing duplicates are merged.
CRM: Lists of companies and deals
HubSpot: Active and static lists (now called segments) for contacts, companies, deals and custom objects (https://knowledge.hubspot.com/segments/create-active-or-static-lists). Today: crm_list_members references contacts only (lib/schema/baseline.ts:227-231); lib/crm.ts:2832-3040. HubSpot segments exist for contacts, companies and deals; account-based teams keep company lists. First scope: an entity on the list (the column exists) honoured by members, dynamic filters and the list page, for companies and deals.
CRM: Export the current view with its columns and filters, to CSV or Excel
HubSpot: Export a view or all records to CSV, XLS or XLSX, choosing the view columns or all properties and including associations (https://knowledge.hubspot.com/import-and-export/export-records). Today: app/api/crm/export/route.ts:46-68 fixed COLUMNS per entity, csv or json. The export today has a fixed column list per object. First scope: export what the table shows (filters, columns incl. custom properties) or all properties, in CSV or XLSX, with the associated record names.
CRM: Import Excel files and several objects in one file, with an import history and undo
HubSpot: Import one or several objects from a file, map columns to properties, dedupe by email, domain or Record ID, and associate the objects of each row (https://knowledge.hubspot.com/import-and-export/import-objects). Today: lib/crm-import.ts:97 one entity per import; :10 and :58-92 company and contact link columns; :439, :514 dedupe. Teams leaving HubSpot export XLSX files with contacts and their companies or deals in one sheet. First scope: accept .xlsx, let one file create contacts plus companies plus deals with the associations of each row, add an Imports page with the result of each import, a downloadable error file and Delete imported records. Also covers the parity rows import-xlsx, import-history.
CRM: Select every record that matches the filter for bulk actions
HubSpot: Select rows or all records in the view, then Assign, Edit any property, Delete, Create tasks, Add to list, Enroll in sequence (https://knowledge.hubspot.com/records/bulk-edit-records). Today: lib/crm-bulk-fields.ts:10-14 BULK_FIELDS; components/crm/ContactsTable.tsx:120-137 enroll; no select-all-matching found in ContactsTable.tsx or BulkEditControls.tsx. Reassigning 2,000 contacts today means 20 pages of selecting. First scope: a Select all N records link that sends the filter to the bulk route instead of ids, with a confirmation that shows the count, plus bulk task creation.
CRM: Add is known, is any of, relative dates and AND/OR groups to filters
HubSpot: Operators depend on the property type: is any of, is none of, contains, is known, is unknown, is between, is more or less than X days ago, rolling date ranges and more (https://knowledge.hubspot.com/segments/determine-filter-criteria). Today: lib/crm.ts:159-162 TEXT_OPS, NUMBER_OPS, SELECT_OPS, DATE_OPS; :204-300 buildRuleCondition. Views and lists copied from HubSpot use these operators all the time (no owner, lead status is any of, last activity more than 30 days ago). First scope: extend buildRuleCondition with known and unknown, any of and none of, not contains, greater or equal, relative date ranges, then OR between groups of AND rules, then filters on the primary company of a contact and the contacts of a deal. Also covers the parity rows filter-groups, filter-associated.
CRM: Make saved views hold columns and sort, update in place and share with a team
HubSpot: Saved views keep filters, columns and sort; they can be updated, cloned, pinned as tabs and set as default (https://knowledge.hubspot.com/records/create-and-manage-saved-views). Today: lib/crm.ts:3050-3080 create, list and delete only; app/api/crm/views/route.ts has no update; lib/schema/baseline.ts:244-251. A HubSpot team rebuilds its views on day one of a migration. First scope: store columns, column order and sort in the view (server side), an Update view action, a visibility of private, team or everyone, and drag to reorder columns. Also covers the parity rows view-columns, view-sharing.
CRM: Load the whole timeline and filter it by type, person and text
HubSpot: The timeline loads older activity as you scroll, grouped by month, with no cap (https://knowledge.hubspot.com/records/filter-activities-on-a-record-timeline). Today: app/[locale]/app/crm/contacts/[id]/page.tsx:255-256 limit: 100; no load-more in components/crm/Timeline.tsx. After a migration records carry years of history and the page stops at the latest 100 items. First scope: cursor pagination in listTimeline with a load-more at the end, a multi-select of activity types, a filter by person or team, a search box, and formatting (bold, lists, links) in notes. Also covers the parity rows timeline-filters, rich-text-notes.
CRM: Rebuild the record page in three columns (properties, activity, associations)
HubSpot: Every record has a left sidebar (actions and About this record properties), a middle column (Overview and Activities tabs) and a right sidebar (association cards and tools) (https://knowledge.hubspot.com/records/work-with-records). Today: components/crm/recordTab.ts:1 RECORD_TABS; app/[locale]/app/crm/contacts/[id]/page.tsx:140-246 header, :263-380 activity tab with aside. This is the page a sales rep lives in. First scope: on contacts, companies and deals at desktop width, a left column with the record name, the action buttons and the editable About properties (the existing PropertiesPanel in compact mode), the timeline in the middle and association cards (companies, deals, contacts, lists, files) on the right; at 390px the columns collapse to the current tabs. Also covers the parity rows record-sidebar-inline-edit.
Cal: Sell meeting packages that bundle several sessions into one payment
Calendly added Meeting Packages and Payment Links in beta as part of its commerce tools (https://community.calendly.com/featured-tips-tricks-32/change-log-march-13-2026-5463). We charge per event type through Stripe (lib/calendar-payments.ts) and each booking is paid on its own. Coaches and consultants who sell five sessions at a price cannot do it without a manual invoice. First scope: a package with a session count and price, a payment link, and a credit counter that is spent as the invitee books each session. Priority: medium. Effort: M.
Cal: Challenge suspicious public bookings with a captcha
Calendly shows invitees a Captcha when a booking attempt looks suspicious (https://community.calendly.com/featured-tips-tricks-32/change-log-march-13-2026-5463). Our public booking endpoint at app/api/calendar/bookings/route.ts has a rate limit but no bot challenge, so a team that shares its page widely can get junk bookings that hold real slots and trigger emails. First scope: a workspace toggle that adds a lightweight proof-of-work or Turnstile challenge to the booking form after repeated attempts from one IP, and logs the blocked attempts in settings. Priority: medium. Effort: M.
Rec: Ask someone for a video with a request link
Loom lets you share a request link so a customer or teammate can record a video straight into your library, without an account in some plans (https://support.atlassian.com/loom/). Rec only records by signed-in members (app/[locale]/app/rec/new), so collecting a bug repro or a client walkthrough means asking them to send a file. First scope: a request link per folder with a prompt and a max length, an anonymous recorder page that uploads through the existing chunked upload (lib/rec-upload.ts), and a Requests entry in the library listing what came in.
Rec: Let people mention teammates in comments and notify them
Loom comments support @mentions that notify the person tagged (https://support.atlassian.com/loom/). Rec comments (components/rec/RecComments.tsx, lib/rec.ts) are plain text with a timestamp and replies, and only the video author is notified. A reviewer who spots a problem at 2:10 has to leave the app to tell the right person. First scope: typing @ in the comment box lists workspace members, the mention is stored with the comment and creates an in-app notification linking to that moment.
Rec: Notify the team in Slack or by webhook when a video is viewed or commented
Loom sends view, comment and reaction notifications and posts video previews to Slack (https://support.atlassian.com/loom/). Rec notifies the author in the app (createNotification in lib/rec-share.ts) and has a partial Slack row, but there is no outgoing webhook on events, so a team cannot route 'client watched the proposal' into Flow or their own tools. First scope: a workspace setting with one webhook URL and a signed JSON payload for video.viewed, comment.created and reaction.created, with a Send test button and a delivery log of the last 20 calls.
Rec: Blur the background of the camera bubble while recording
Loom offers virtual backgrounds, camera frames and background blur on paid plans (https://support.atlassian.com/loom/). Rec draws the camera bubble into the recording (components/rec/Recorder.tsx, lib/rec-controls.ts) with no effect options, and the parity table lists camera-effects as missing (content/feature-parity/rec.ts). People recording from home or a shared office care about what shows behind them, and it is the most asked effect. First scope: a Blur toggle in the camera settings of the recorder using a canvas filter or the browser's background segmentation when available, off by default and hidden when unsupported.
Rec: Start embedded videos at a given second and autoplay from the embed URL
Loom lets you add URL parameters to the embed code to start a video at a timestamp, autoplay it or hide player elements (https://support.atlassian.com/loom/docs/make-embedded-videos-autoplay-or-play-at-a-specific-time/). Rec now opens the share page at ?t= (components/rec/RecWatch.tsx), but the embed page (components/rec/RecEmbed.tsx, app/[locale]/v/[share]/embed) ignores any parameter and the embed code in RecShareDialog has no options. Teams paste a video into a help article or a Notion page and want it to begin at the step that matters. First scope: honor t and autoplay (muted, as browsers require) on the embed route and add a Start at field to the embed dialog.
Forms: Pass the answers to the redirect URL as query parameters
Typeform redirects after submit and can append answers and hidden fields to the URL using variables (https://www.typeform.com/help/a/redirect-on-completion-360029618711/). Forms redirects to a fixed https URL per ending (redirectUrl in lib/forms/types.ts) and drops the answers. Teams use this to prefill a Calendly, Cal or payment page with the name and email just collected. First scope: placeholders like {{email}} in the redirect URL, piped with the form's own pipe function and URL encoded.
Forms: Keep chosen options pinned while the rest are randomized
Typeform cannot randomize choices while keeping an option such as None of the above in place, and its community asks for it repeatedly (https://community.typeform.com/build-your-typeform-7/how-do-i-randomize-choices-while-fixing-position-of-certain-options-4957). Forms now randomizes the order of choices per respondent, but the whole list moves. Surveys usually need Other or I don't know to stay last. First scope: a pinned flag on each choice that keeps it at the end of the shuffled list, editable in the options list of the builder and covered by the shuffle unit test.
Forms: Add an Address question with street, city, region and country
Typeform Contact info and Address blocks collect several fields in one screen (https://www.typeform.com/help/a/address-360029572971/). Forms has one field per question, so asking for a shipping address takes five questions with five Enter presses. Delivery, event and hiring forms ask for addresses all the time. First scope: an address question type with street, city, region, postal code and country fields, required per field, saved as an object and exported as separate CSV columns.
Forms: Email a copy of the answers to the respondent
Typeform sends respondents an email with a recap of their answers when the form has an email question and the setting is on (https://www.typeform.com/help/a/send-a-notification-to-respondents-360029603912/). Forms only notifies the team (notifyPeople and notifyEmails in lib/forms/types.ts) and shows an ending screen. Booking, application and order forms need a receipt the person can keep. First scope: a settings toggle that picks one email question and sends a plain message with the answers piped in, through the existing mail sender with its fake flag.
Forms: Add picture choice questions with an image per option
Typeform has a Picture choice question where each option is an image with a label (https://www.typeform.com/help/a/picture-choice-360029568931/). Forms has text-only options in lib/forms/types.ts and components/forms/QuestionInputs.tsx. Product feedback, event and menu forms read much better with images. First scope: an optional imageUrl on each choice, uploaded through the existing form file route, rendered as a grid of cards that keep the keyboard letters.
Forms: Add an Other option with a free text field to choice questions
Typeform multiple choice questions can add an Other option that opens a text box, and Tally keeps it pinned last when choices are randomized (https://tally.so/help/how-to-create-an-other-option). Our choice, checkbox and dropdown questions only accept the fixed list in lib/forms/types.ts, so a respondent whose answer is not listed has nowhere to write it. Small teams run into this on every survey. First scope: an allowOther flag on choice and multi_choice that shows a last option with a text input, stored in the answer and shown in the responses table, summary and CSV.
Robots: Pause a run for a human decision and resume from the app
UiPath Action Center lets a robot hand a task to a person and continue after they respond (https://docs.uipath.com/action-center/automation-cloud/latest/user-guide/about-action-center). Robots run start to finish with no way to wait for approval, for example before submitting a form. First scope: an Ask for approval step that stops the run in a waiting status, notifies the workspace, and continues or fails when someone answers in the run view.
Robots: Show a run timeline with the duration of each step
UiPath Orchestrator shows execution logs per job with timestamps for each activity (https://docs.uipath.com/orchestrator/automation-cloud/latest/user-guide/logs). Our run detail lists steps and status (components/robots/RunsPanel.tsx) but not how long each took. A team tuning a slow robot cannot see which step eats the time budget. First scope: store started and finished times per step in the run log and show the duration next to each step.
Robots: Retry a group of steps together with a limit and delay
UiPath's Retry Scope reruns a whole sequence until a condition passes (https://docs.uipath.com/activities/other/latest/ui-automation/activities-index). Our retries apply to one step at a time (retries and retryDelayMs in lib/robots-schema.ts), so a flow like open menu, click, wait for result cannot be retried as a unit. First scope: a group step with up to 5 attempts that reruns its children from the first one and logs each attempt.
Robots: Add a Hover Text step that finds text and hovers over it
UiPath has a Hover activity and a Hover Text activity that searches a string inside an element and hovers it (https://docs.uipath.com/activities/other/latest/ui-automation/hover-text). Our Click step takes a selector or visible text, but nothing hovers (lib/robots-interpreter.ts). Menus that open on hover block small teams from automating many admin panels. First scope: a hover step with selector or visible text, plus a short wait for the menu to render.
Robots: Add a Toggle option to the Check or uncheck step
UiPath's Check/Uncheck activity offers Check, Uncheck and Toggle (switch the current state) (https://docs.uipath.com/activities/other/latest/ui-automation/activities-index). Our Check or uncheck step (lib/robots-schema.ts) only sets a final state of checked or unchecked. Toggle matters for switches that a site renders as a custom button where the current state is not readable. First scope: a third state value that clicks once and logs the result.
Board: Let an admin set a board to view-only for everyone
Miro can restrict a board to view access for most people while owners keep editing (https://help.miro.com/hc/en-us/sections/360003285093-Working-on-the-Board). Board has per-person access levels in lib/board-access.ts, but no single switch to freeze a finished board such as a signed-off retro or a workshop result. Without it, a stray drag by a teammate changes a record the team wanted to keep. First scope: a read-only toggle in the share panel that makes requireBoardAccess treat edit as view for non-owners, with a visible banner in the editor and an audit log entry.
Board: Add a keyboard shortcut sheet for the sticky, table and presentation tools
Miro opens a shortcut reference with F1 and lists every hotkey (https://help.miro.com/hc/en-us/sections/360003285093-Working-on-the-Board). Board relies on the stock Excalidraw help dialog, which does not mention our own tools: sticky notes, tables, reactions, voting, follow mode and present (components/board/BoardCanvasTools.tsx, BoardPresent.tsx). Shortcuts we add are invisible unless someone reads the changelog. First scope: a '?' dialog in the editor chrome that lists the Board-specific keys in en, es and pt and opens from Cmd+K.
Board: Save a board as a workspace template and start new boards from it
Miro lets teams keep their own templates next to the built-in library (https://help.miro.com/hc/en-us/sections/360003285093-Working-on-the-Board). Board ships a fixed gallery in lib/board-templates.ts and a duplicate action, but a team cannot mark its retro or planning board as a reusable template. Small teams repeat the same ritual every sprint and today they duplicate an old board and clear it by hand. First scope: a 'Save as template' action in the board menu that stores a copy flagged as template, shown in the gallery under a 'Your templates' heading and hidden from the main list.
Notes: Let a public link expire on a date and require a password
Evernote public links stay open until revoked, and its help center documents no expiry (https://help.evernote.com/hc/articles/208313588). Notes has a public read link (share_token in lib/notes-shares.ts) that works until someone turns it off. For a small team sharing client notes, a link that expires after a week and an optional password is a safer default than remembering to revoke. First scope: optional expires_at on the share, checked on the public route, with a picker in the share popover; password protection as a second step.
Notes: Parse natural-language dates like 'next Friday' when inserting a date
Evernote 11.26 added natural-language date insertion in the editor (https://releasebot.io/updates/evernote). Notes inherits the Docs date mention with today, tomorrow and a calendar picker (components/docs/DocsEditor.tsx), but typing a phrase such as 'next friday' or 'en 3 dias' does not resolve to a date. Dates in meeting notes and checklists are written fast, and the same helper could feed checklist due dates and reminders. First scope: a pure parser for en, es and pt with unit tests, used by the date mention input.
Notes: Transcribe uploaded video files with speaker labels
Evernote 11.26 added video transcription with speaker recognition (https://releasebot.io/updates/evernote). Notes transcribes recorded or uploaded audio through POST /api/notes/[id]/transcribe and already labels speakers, but a video file attached to a note cannot be transcribed. Teams paste screen recordings of calls into notes, so this closes the gap. First scope: accept video/mp4 and video/webm in the transcribe route by extracting the audio track, keeping the current size limit and AI quota check.
Notes: Show the length of audio recordings and uploaded audio files
Evernote 11.25 shows the recording length on audio and added automatic transcription of uploaded audio in 11.23 (https://releasebot.io/updates/evernote). Notes records and transcribes audio (AudioRecordButton in components/notes/NoteEditorPane.tsx) but the audio block gives no total duration before playing it, and the transcript heading carries no timestamps. Knowing that a clip is 42 minutes before spending AI quota on it matters for a small team. First scope: read the duration from the audio metadata, show it on the block and in the transcribe confirmation.
Notes: Add highlighter colors to the text toolbar
Evernote 11.25 added highlighter color options (https://releasebot.io/updates/evernote). The Notes editor is the Docs block editor and Notes has no color highlight in its toolbar; the only highlight today is the search-term mark in components/notes/NoteEditorPane.tsx. Highlighting key lines is a core habit when a small team uses notes for meeting minutes and research. First scope: five token-based highlight colors on selected text that work in light and dark mode, kept in Markdown and ENEX export as plain text.
- 0
Lines: Phone number masking for members
Periskope hides customer phone numbers from some roles. First scope: a workspace setting that shows only the last 4 digits to members in the inbox, lists, exports and API. Priority: low. Effort: S.
- 0
Lines: Saved chat lists as broadcast audiences
Periskope saves filtered chat lists and uses them as broadcast audiences. First scope: save the current inbox filters as a named list in the sidebar, and pick a list as recipients in a new broadcast. Priority: low. Effort: S.
- 0
Lines: More webhook events and group endpoints in the public API
Lines emits message received and ticket created or updated, and the API covers chats, messages and tickets. Periskope has 20 webhook events and group endpoints (create, members, admins, invite). First scope: chat.created, message.ack, ticket.closed, and /api/v1/groups/groups endpoints for create and members. Priority: medium. Effort: M.
- 0
Lines: Edit, delete, forward and pin messages
Periskope edits, deletes, forwards and pins WhatsApp messages from the inbox. The WAHA and Evolution APIs support them. First scope: the message actions menu plus the provider calls. Priority: low. Effort: M.
- 0
Lines: AI flags only real questions as unanswered
Lines flags a chat by time without reply. Periskope uses AI to tell questions and requests from comments like "thanks", so fewer chats are flagged. First scope: a cheap classifier on each inbound message (lib/ai.ts, off when AI is off) that skips awaiting for non-questions. Priority: medium. Effort: M.
- 0
Lines: @mentions in internal notes
Notes in a chat are visible to the team but do not notify anyone. First scope: @person in a note sends an in-app notification (the bell) and an email per the person's Track notification settings. Priority: medium. Effort: S.
- 0
Lines: Full group settings (photo, admin-only messages, approval to join)
Lines changes name and description. Periskope also sets the photo, who can send messages, who can edit info, join approval and disappearing messages. WAHA and Evolution API expose these calls. First scope: the toggles in the group page, in bulk from the list. Priority: medium. Effort: M.
- 0
Lines: Group templates with preset members and settings
Periskope creates groups from templates with the account team already inside. First scope: save a template (name pattern, members, admins, description, labels) and use it in New group and the CSV import. Priority: medium. Effort: S.
- 0
Lines: Recurring scheduled messages
Periskope repeats scheduled messages daily, weekly or monthly with an end date. Lines schedules one-off messages and broadcasts. First scope: a repeat rule on a scheduled message that creates the next send after each run (cron claim per row). Priority: medium. Effort: S.
- 0
Lines: Members see only the chats with their labels
Periskope limits what a Member sees by label, which is how teams keep private chats per account owner. First scope: a label list per person in Lines settings; members only list, search and open chats with one of their labels; admins see everything. Priority: high. Effort: M.
- 0
Lines: Round-robin assignment of new chats
Periskope assigns new chats by rules (round-robin among online agents). Today Lines assigns by hand. First scope: an on/off rule per number that rotates new chats among chosen people, skipping people marked away. Priority: high. Effort: M.
- 0
Lines: WhatsApp team inbox with group management, tickets and broadcasts
Periskope (https://periskope.app) gives B2B teams one inbox for every WhatsApp customer group. Build it in the suite: numbers by the official Cloud API or a WAHA / Evolution API gateway, shared inbox for chats and groups, bulk group management, tickets by emoji, SLAs, broadcasts, analytics and the group on its CRM record.
- 0
Chat: Sort the inbox by newest, oldest or longest wait
Intercom lets the team sort the inbox by Waiting since, ascending, so the visitor who has waited longest comes first (https://www.intercom.com/help/en/articles/6989006-inbox-sorting). Our inbox was fixed to latest activity (lib/chat-data.ts). A small team with no SLA policy had no way to see who was overdue. Built: a Sort select in the inbox with newest, oldest and waiting longest, where waiting counts from the visitor's first message after the last human or bot reply and ignores internal notes.
Chat: Combine priority and waiting time in one inbox sort
Intercom users ask for sorting that puts urgent conversations first and then the longest waiting, because the two sorts cannot be combined today (https://community.intercom.com/ideas/combined-sorting-for-priority-and-non-priority-conversations-13151). Our inbox now sorts by newest, oldest or waiting longest (lib/chat-data.ts conversationOrderBy) but priority is only a filter. First scope: a fourth sort option, Priority then waiting, ordering urgent, high, medium and low and breaking ties by waiting since.
Chat: Show how long each visitor has been waiting in the inbox rows
Intercom shows a waiting-since time on each conversation row so the team sees at a glance who is overdue (https://www.intercom.com/help/en/articles/6989006-inbox-sorting). Our rows in components/chat/ConversationList.tsx show the last message and SLA badges only when a policy exists, so a team without SLA policies cannot see who waits longest. First scope: a short relative label such as 12m or 3h on rows where the last message is from the visitor, computed from the same waiting-since rule used for the Waiting longest sort.
Chat: Schedule an agent reply to send later
Intercom lets teammates schedule a message to go out at a chosen time, useful when replying across time zones (Intercom help center, inbox composer). Our composer in components/chat/ConversationClient.tsx only sends immediately or saves an internal note. A team in Buenos Aires answering a visitor in Madrid ends up writing at the wrong hour or forgetting the reply. First scope: a Send later option with a date and time, a pending row in the conversation that can be edited or cancelled, and delivery from the minute cron with a compare-and-set claim.
Chat: Show expected reply time in the widget header
Intercom's Messenger can show a reply time expectation such as a few minutes or a day (Intercom help center, Messenger settings). Our widget (app/api/chat/widget.js/route.ts) shows office hours state but never says how long a visitor should wait. Visitors who write outside the first minutes assume nobody is there and leave. First scope: a per-widget choice among a few fixed options in WidgetSettings, rendered under the widget title in en/es/pt, and hidden when the team is away per office hours.
Chat: Block a visitor or mark a conversation as spam from the inbox
Intercom lets teammates mark a conversation as spam and block the sender so they stop reaching the inbox (Intercom help center, Inbox section). Our inbox has no way to ban a visitor: lib/chat-data.ts and app/api/chat/messages/route.ts accept messages from any visitor key and the only recourse is resolving each conversation. A small team on a public widget gets junk quickly and has nobody to filter it. First scope: a blocked flag on chat_visitors, a Block action in the visitor panel, new messages from that visitor dropped silently by the widget endpoint, and a list of blocked visitors in widget settings with unblock.
- 0
Link: Let search engines index a short link with a per-link indexing switch
Dub exposes a doIndex boolean per link and sends noindex by default (https://dub.co/docs/concepts/links/introduction). Link redirects in app/r/[slug]/route.ts sent no X-Robots-Tag, so short URLs could end up in search results. Built: redirects carry X-Robots-Tag: noindex unless the link turns on doIndex, with a checkbox in EditLinkForm and the field in the API.
Link: Change the destination by date range with time-based redirect rules
Lyncks lets one link change its destination automatically based on the time or date you set (https://www.hunted.space/product/lyncks). Link targeting covers country, device and A/B variants in lib/link-targeting.ts, with nothing based on time. A restaurant or a webinar team reuses one printed QR code and swaps the destination by hand each time. First scope: an ordered list of date range rules in the targeting JSON, checked before A/B, with the base URL as the fallback.
Link: Email a weekly digest with total clicks and the top links
Linkbreakers sends weekly and monthly digests with total visits, top country and top link (https://linkbreakers.com/help/article/weekly-and-monthly-email-digests). Link has analytics pages and CSV export (app/[locale]/app/links/analytics) but nothing reaches the owner unless they open the app. A small team with a few campaign links wants one Monday email with clicks against the previous week. First scope: an hourly tick job that sends one email per workspace owner who opted in from settings, built from the queries in lib/link-analytics.ts.
Link: Check destinations daily and flag links that return 404 or 5xx
Redirect and URL monitors such as Pagecrawl check a destination on a schedule and alert when it breaks, shortened links included (https://pagecrawl.io/tools/monitor-website-redirects-url-changes.html). Link never fetches a destination after creation: lib/links.ts only validates the URL syntax, so a printed QR can point at a dead page for weeks. First scope: a daily cron under app/api/cron that sends a HEAD request for links with clicks in the last 30 days, stores last_check_status and last_checked_at, shows a warning in components/link/LinksTable.tsx and the link detail, and emits a link.broken webhook event.
Link: Add a start date so a link only goes live from a set time
The Expired Link Redirection plugin schedules a link between two dates with a before, during and after URL (https://wordpress.org/plugins/expired-link-redirection/), and Lyncks changes a link destination by date (https://www.hunted.space/product/lyncks). Link only supports an end: expires_at and max_clicks in lib/links.ts, enforced in app/r/[slug]/route.ts. A small team printing a QR code ahead of a launch has to create the link on launch day. First scope: a nullable starts_at column, a datetime field next to expiry in components/link/EditLinkForm.tsx, and the redirect answering with the link fallback or a not-live-yet reason on /gone before that time.
CRM: Propose stakeholders and pain points from deal notes and call transcripts
HubSpot now scans notes and call or meeting transcripts on a deal and adds stakeholders and pain points as structured signals on the associated company record (https://martech.org/hubspots-june-2026-updates-agents-get-more-access-revenue-hub-starts-connecting-the-dots/). We have an AI summary card and call summaries in lib/crm-ai.ts, but their output stays as text on the timeline and nothing is written to properties or associations. A first scope is a button on the deal that proposes people mentioned in notes as contacts with a role label and a short list of pain points, applied only after the user confirms each one.
CRM: Add a unique-value option to custom properties to block duplicate IDs
HubSpot custom properties can be marked as unique identifiers so two records cannot share the same value, which keeps external IDs and tax numbers clean (https://knowledge.hubspot.com/properties/create-and-edit-properties). Our custom field definitions in lib/crm-custom-fields.ts have no uniqueness rule, so the same external ID can be saved on two contacts and only the duplicate finder catches it later. A first scope is a unique flag on text and number fields that rejects the save in the UI, the CSV import and API v1 with a message naming the record that already has the value.
CRM: Group custom properties into named sections
HubSpot organizes properties into property groups that appear as sections in forms, record pages and imports (https://knowledge.hubspot.com/properties/create-and-edit-property-groups). Our custom fields in crm_custom_fields have a name, label, type and position with no grouping, so a workspace with 20 fields gets one flat list in components/crm/CustomFieldsPanel.tsx. A first scope is an optional group name on each field, shown as collapsible sections on the record page and the settings screen, with a migration that keeps existing fields ungrouped.
CRM: Let admins arrange the cards and properties of the record sidebar
HubSpot admins customize record layouts under Settings > Objects > Record Customization, adding cards and choosing which properties show in the About card, and each user can reorder cards (https://knowledge.hubspot.com/object-settings/customize-records). In our CRM the contact, company and deal pages show a fixed set of panels under app/[locale]/app/crm. Teams with many custom fields end up scrolling past properties they never use. A first scope is a per-object list of visible properties and their order for the side panel, stored per workspace, with a reset to default.
CRM: Add calculated rollup properties (count, sum, min, max) from associated records
HubSpot lets admins create rollup properties that compute count, sum, min, max or average over associated records, optionally filtered by association label, plus calculation properties built from an equation or time-between logic (https://knowledge.hubspot.com/properties/create-calculation-properties). Our custom fields in lib/crm-custom-fields.ts are all typed inputs entered by hand, and companyRollup in lib/crm.ts is a fixed set of totals. A small team wants 'total won revenue' on a company or 'open deals' on a contact as a column and a filter without building a report. A first scope is a read-only property type on companies and contacts with count and sum over associated deals, recomputed when a deal changes, usable as a column, sort and filter.
- 0
Docs: Show word count and reading time in the page menu
Notion shows a word count at the bottom of the page menu (https://bullet.so/blog/how-to-check-word-count-in-notion/). Docs showed nothing before this change. The page menu in components/docs/DocsEditor.tsx now lists words, characters and reading time, computed by lib/docs-stats.ts.
Docs: Add a Markdown endpoint to read and write a page body in the public API
Notion's API returns and accepts page content as Markdown through GET /pages/{id}/markdown (https://matthiasfrank.de/en/notion-updates/api-markdown-and-transcript-support/). Docs has export to Markdown in the UI (lib/docs-markdown.ts) and a public API (lib/docs-v1.ts), but scripts and AI tools must handle the block JSON. First scope: GET and PUT /api/v1/docs/pages/{id}/markdown behind docs:read and docs:write, reusing the existing exporter and the paste-Markdown parser.
Docs: Show sibling pages when hovering a breadcrumb segment
Notion's April 2026 release lets you hover any breadcrumb item to see its sibling pages, including pages in databases (https://www.notion.com/releases/2026-04-24). Docs renders the ancestor trail as plain links (components/docs/DocsEditor.tsx, docs-breadcrumb). Moving sideways in a wiki then needs the sidebar. First scope: a menu on each segment that lists the siblings from the page tree, with keyboard support.
Docs: Present a page as slides from a heading outline
Notion graduated presentation mode out of beta on paid plans in 3.4 (https://www.notion.com/releases). Docs has no full-screen presentation of a page. Teams run weekly reviews from a page and would not need a second tool. First scope: a Present action in the page menu that splits blocks at each H1 or divider, steps with arrow keys, and exits with Esc.
Docs: Add a Tabs block to split a long page into tabs
Notion 3.4 added a tabs block for organizing a page into sections (https://www.notion.com/releases). Docs has columns and toggles (lib/docs-columns.ts, components/docs/DocsBlocksView.tsx) but no way to show one section at a time inside a page. Small teams use it for project hubs with Overview, Decisions and Links. First scope: a tabs block with named tabs that each hold child blocks, switched with the keyboard, rendered in the read-only and published views.
Docs: Find text inside the open page with Cmd+F and accent-insensitive matches
Notion 3.2 ships in-page find that ignores accents, so searching 'cafe' matches 'café' (https://www.notion.com/releases/2026-01-20). Docs has a Cmd+P page switcher and content search across pages (lib/docs-search.ts), but nothing that highlights matches inside the page being edited, so the browser find misses text in collapsed toggles. This matters for long meeting notes and specs. First scope: a find bar on Cmd+F that highlights matches, expands toggles, steps with Enter and Shift+Enter, and reuses normalizeSearch for accents.
Flow: Add a Convert to file step for CSV and JSON attachments
n8n Convert to File and Extract from File turn rows into CSV or JSON and back (https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.converttofile/). Flow has no binary or file data between steps: action.email sends text only (lib/flow-schema.ts). Weekly report flows end up pasting rows into the email body. First scope: build a CSV string from a list and attach it to the email step.
Flow: Add a Compare datasets step to find new, changed and removed rows
n8n Compare Datasets returns items only in A, only in B, equal and different (https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.comparedatasets/). Flow has no way to diff two lists without writing code in data.code (lib/flow-schema.ts). Syncing a Sheet with CRM contacts is a common first flow for small teams. First scope: two lists, a key field and four outputs.
Flow: Add a Digest step that collects items and releases them on a schedule
Zapier Digest collects entries over time and releases them as one summary (https://zapier.com/apps/digest/integrations). Flow can approximate it with a data table and a schedule flow (lib/flow-tables.ts) but needs two flows and manual cleanup. A Digest step would append to a named bucket and a schedule trigger would read and clear it. This matters for teams that want one daily email instead of one email per event. First scope: append, release and clear on a per-flow bucket.
Flow: Add a Split out step that turns a list field into one item per entry
n8n has Split Out and Aggregate nodes to move between a list and separate items (https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.splitout/). Flow loops over a list with logic.loop (lib/flow-schema.ts) capped at 100 items and with no Delay inside. A step that fans a list into items would let nodes after it run per entry without those limits. First scope: a list template and a field to keep, up to 1,000 items.
Flow: Add a Stop and error step that fails a run with a custom message
n8n has a Stop and Error node (https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.stopanderror/) and Zapier has Paths with a halt option. Flow can only fail through a node error or a Code step (lib/flow-engine.ts), so a branch that should end as failed needs a script. A step with a message template would route into the existing failure alert and error flow. First scope: one field with the message and an optional error code.
- 0
Flow: Sort, limit and remove duplicates from a list in one step
Done. n8n has Sort, Limit and Remove Duplicates nodes (https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.sort/) and Zapier Formatter has Utilities such as pick from list. Flow had only Merge/Aggregate and the Code step (lib/flow-schema.ts, data.merge). The new List step covers sort, limit and dedupe over a list from an earlier node.
Agents: Add a bulk re-run of saved tests across all agents from one screen
Fin lets teams test changes against many past conversations before turning them on (https://www.eesel.ai/blog/intercom-fin-ai-explained). Saved tests exist per agent (components/agents/AgentTests.tsx) but each agent must be opened one by one. First scope: a Tests overview on the agents home that runs all tests for the selected agents and shows pass or fail counts with a link to each failing case.
Agents: Show a preview of which knowledge source answered each reply
Fin shows the content used to produce each answer in the conversation view so admins can fix the source (https://www.intercom.com/help/en/articles/8205718-fin-resolutions). We cite sources to the visitor (lib/agents-semantic.ts) but the conversation view in components/agents/AgentConversations.tsx does not list which chunks were retrieved. First scope: store the retrieved source ids per assistant message and render them under the reply with a link to open the Docs page or file.
Agents: Add custom guidance rules that apply by audience or topic
Intercom Fin has Guidance, short rules that apply only when a condition matches, such as the customer's plan or topic (https://www.intercom.com/help/en/articles/10444459-fin-guidance). Here the only free-form text is one Instructions field plus procedures (lib/agents-procedures.ts), so every rule is always in the prompt. A small team with a paid plan and a free plan needs different answers about refunds. First scope: a list of rules with a condition text and an instruction, matched per message by keywords, and a count of how often each fired in the analytics tab.
Agents: Export and import an agent as a JSON file
Chatbase and Voiceflow let teams move agent configuration between workspaces (https://www.voiceflow.com/docs/documentation/deploy/environments/overview.md). Duplicate now copies an agent inside a workspace (lib/agents.ts duplicateAgent), but there is no file to back up before a big edit or to share a template. First scope: Export button in agent Settings that downloads instructions, knowledge selection, tools, handoff, guardrails and saved tests as JSON without secrets, and an Import button on the agents list that creates a new agent and lists what must be reconnected.
- 0
Notes: Duplicate a note from its menu
Evernote and most note apps let you copy a note to start a variation, for example a weekly template (https://help.evernote.com/hc/articles/209004637). Notes had no duplicate action in components/notes/NoteEditorPane.tsx. First scope: a Duplicate entry in the note menu that creates a copy with a Copy of title, same tags and notebook, and opens it.
Notes: Insert a table of contents block from the note headings
Evernote can build a table of contents note that links to other notes, and Notion-style tools offer an in-page contents block (https://techlearning.com/tl-advisor-blog/9011). Our editor supports headings and [[ ]] links in components/notes/NoteBlocksView.tsx but no contents block. Long notes such as meeting logs and guides are hard to navigate. First scope: a /contents slash command that inserts a block listing the note headings as anchor links and updates as headings change.
Notes: Add a keyboard shortcuts drawer opened with Cmd+/
Evernote shows all shortcuts in a drawer opened with Cmd+/ or Ctrl+/ and lets you remap them (https://help.evernote.com/hc/en-us/articles/34296687388307). In components/notes/NotesApp.tsx the shortcuts N and / exist but are only discoverable through small kbd hints. First scope: a read-only drawer listing the shortcuts of the list and the editor, opened with Cmd+/ and from the sidebar help menu, in the three languages. Remapping can come later.
- 0
Notes: Add a Note info panel with word count, dates and links
Evernote has a Note info command (Ctrl+Shift+I) that shows created and updated dates, size and word count (https://help.evernote.com/hc/en-us/articles/34296687388307). Our editor in components/notes/NoteEditorPane.tsx shows only the save status and has no word or character count. Writers and people drafting posts need the count while typing. First scope: a menu entry and a shortcut that open a small popover with words, characters, reading time, created and updated dates, notebook and source URL.
Notes: Merge several selected notes into one note
Evernote lets you select several notes and merge them into one, keeping the first-selected note on top (https://techlearning.com/tl-advisor-blog/9011, an older guide to check against the current app). Today the list selection in components/notes/NotesApp.tsx offers move, tag, pin and trash through app/api/notes/bulk/route.ts, but not merge. Small teams collect meeting fragments in many notes and want one clean page. First scope: a Merge action in the selection bar that concatenates content in selection order, unions tags, keeps the first note and trashes the others, with the existing version history as the undo.
Board: Add estimation sessions with hidden votes and a reveal button
Miro offers a planning poker and estimation app where everyone picks a card and the facilitator reveals all at once (https://miro.com/templates/planning-poker/). We have dot voting and a shared timer in lib/board-workshop.ts but votes there are visible while people cast them. For sprint planning on Board plus Track, hidden votes avoid anchoring. First scope: a workshop mode with a fixed Fibonacci scale, hidden picks per person, a reveal action for the facilitator and a button that writes the median to the selected sticky.
- 0
Board: Add a table object with rows, columns and paste from a spreadsheet
Miro has a Table item with typed columns that can be filled by pasting from Sheets or Excel (https://help.miro.com/hc/en-us/articles/360017571174). Our canvas is Excalidraw, which has no table element, and nothing in lib/board-templates.ts or components/board creates one. Comparison matrices and prioritization grids are a common reason teams open a whiteboard. First scope: paste of tab-separated text creates a grid of bound text cells aligned on the canvas, grouped so it moves as one.
Board: Add emoji reactions that float on the canvas for workshops
Miro has reactions that people trigger with a shortcut and that float up next to their cursor, used for quick feedback during a session (https://help.miro.com/hc/en-us/articles/360017731113). We have live cursors and presence (components/board/BoardPresence.tsx) but no way to react without adding a comment or a sticky. First scope: five fixed emojis in the live layer, sent over the existing live channel in lib/board-live.ts, shown for three seconds and never saved in the scene.
Board: Let people request access from a private board link
Miro shows 'Request access' to anyone who opens a board link they cannot see, and notifies the owner (https://help.miro.com/hc/en-us/articles/360017730693). Here a user without access gets a plain error page from app/[locale]/app/board/[id]. The result is a Chat message asking the owner to share it manually. First scope: a request button on the no-access page that creates a notification for the owner, who can grant viewer or editor from BoardAccessPanel in one click.
- 0
Board: Star boards and pin them at the top of the list and sidebar
Miro has Starred boards that stay pinned on the home screen (https://help.miro.com/hc/en-us/articles/360017730773). Our board list in components/board/BoardList.tsx sorts by update date only, with no per-user favorite. Teams that keep one living roadmap board plus many workshop boards lose it under the new ones. First scope: a per-user star toggle on each row and in the editor header, stored in a small table, with starred boards listed first and in the sidebar.
Board: Copy a link that opens the board zoomed on a frame or selection
Miro lets you copy a link to a single object or frame so teammates land on that part of a large board (https://help.miro.com/hc/en-us/articles/360017730533). Our boards open at the last saved viewport and there is no deep link parameter, as lib/board-urls.ts only builds board-level URLs. In a retro or planning board with ten frames, people paste 'see the third column' in chat instead of a link. First scope: a 'Copy link' action on a selected frame or element that adds ?focus=<elementId>, and the editor and viewer scroll to it on load.
- 0
Robots: Add a Check or uncheck step for checkboxes and radio buttons
UiPath has a dedicated Check/Uncheck activity that sets a checkbox to a known state instead of toggling it (https://docs.uipath.com/activities/other/latest/ui-automation/). Our Click step toggles, so a robot that runs on a page where the box starts checked flips it the wrong way. First scope: a check step with a checked/unchecked choice that uses the page's real state, and a recorder mapping from checkbox clicks to this step.
Robots: Dismiss cookie banners and dialogs automatically before each step
UiPath handles unexpected popups with Pop-up handlers triggered during execution (https://docs.uipath.com/activities/other/latest/ui-automation/). Robots here fail or click the wrong element when a consent banner covers the page, and the user must record a manual click step for every site. For a small team that scrapes many sites this is the most common flaky step. First scope: a robot setting that clicks known consent buttons and auto-accepts JavaScript alert dialogs, with a note in the run log when it fired.
Robots: Switch between tabs and handle popups opened by a click
UiPath's Attach Browser and Use Browser scopes let a workflow move between windows and tabs (https://docs.uipath.com/activities/other/latest/ui-automation/). Our interpreter runs against a single page (lib/robots-interpreter.ts), so a link with target=_blank or a login popup leaves the robot on the old page. This blocks a lot of real portals. First scope: a `switch_tab` step (newest, by URL contains, back to first) and an option on Click to follow the new tab.
Robots: Add an Assert step that fails the run when a value or element is wrong
UiPath ships Verify Expression and Check App State activities to validate a page before continuing (https://docs.uipath.com/activities/other/latest/ui-automation/). We have If exists, which can fail on a missing element, but nothing that checks a value such as 'row count is greater than 0' or 'text contains Welcome'. Robots that silently return empty tables look like successes in the run history. First scope: an assert step with element exists, text contains and variable comparisons, and a clear error in the run log.
Robots: Add a Set variable step with text transforms (trim, regex, replace, number, date)
UiPath's Assign and string activities let a workflow reshape values between steps (https://docs.uipath.com/activities/other/latest/ui-automation/). Today our robots can only write variables through Extract or Evaluate with custom JavaScript (lib/robots-interpreter.ts), so cleaning a price or building a URL needs code. A small team wants '$1,299.00' to become 1299 without writing a script. First scope: a set_var step with a template and a list of transforms, usable in later steps as {{vars.name}}.
Robots: Add hover, double click and right click actions to the Click step
UiPath's Click activity exposes click type (single, double) and mouse button, and hover is a separate Hover activity (https://docs.uipath.com/activities/other/latest/ui-automation/). Our Click step in lib/robots-schema.ts only does a single left click, so menus that open on hover and context menus cannot be automated. Small teams hit this on dashboards with hover menus. First scope: a `mode` option (click, double, right, hover) on the click step, recorded by the recorder when it sees dblclick or contextmenu events.
CRM: Separate the permission to create records from the permission to edit
HubSpot admins can allow a role to edit records without letting it create new ones (https://knowledge.hubspot.com/user-management/hubspot-user-permissions-guide). Our visibility rules in lib/crm-visibility.ts and lib/crm-field-perms.ts control what a person sees and which fields they change, not whether they can add records. This stops interns or guests from filling the CRM with duplicates. First scope: a per-role create toggle for contacts, companies and deals checked by the POST routes under app/api/crm.
CRM: Add formula properties that calculate from other properties
HubSpot calculation properties compute a number, date or text from other properties on the record (https://knowledge.hubspot.com/properties/create-calculation-properties). We only have stored custom fields in lib/crm-custom-fields.ts. A small team uses this for margin, days to close or weighted amount without a report. First scope: a formula field type with arithmetic over number and currency fields, computed on read, usable in table columns and filters.
CRM: Set default values for properties on new records
HubSpot admins can set a default value per property that applies on manual creation, workflows and the API (https://knowledge.hubspot.com/object-settings/set-default-property-values). Our custom fields have no default, so every new contact starts blank and people forget the same field. First scope: a default per custom field and per core property such as source or priority, applied in createContact, createCompany and createDeal in lib/crm.ts when the value is missing.
CRM: Make custom properties required on the create form
HubSpot properties can be marked required so a record cannot be created without them (https://knowledge.hubspot.com/object-settings/set-required-properties-for-records). Our custom fields in lib/crm-custom-fields.ts only support types and options, and stage rules in lib/crm-stage-rules.ts cover deals entering a stage, not creation. For a small team this keeps owner, source or region from being empty in reports. First scope: a required flag per custom field, enforced in the create forms and the v1 API with a clear error.
- 0
CRM: Clone a deal with its associations from the record menu
HubSpot lets you clone a deal, contact or company from the record actions menu (https://knowledge.hubspot.com/records/clone-records). We had no way to copy a record, so repeat deals meant retyping every property. First scope: Clone deal in the More menu at app/[locale]/app/crm/deals/[id]/page.tsx, copying properties, contacts and companies but not notes, tasks or line items; contacts and companies come later.
- 0
Track: Preload a default issue template per project in quick create
Linear lets a team mark a template as the default so every new issue starts from it (https://linear.app/docs/issue-templates). Track had templates per project (lib/track-templates.ts) that had to be picked by hand each time. A small team that files most issues in one shape, such as bug reports, repeats that pick on every issue. Built: a Default checkbox on the template form, one default per project, and quick create fills it in on an empty draft.
Track: Send a saved view digest to Slack or email
Linear lets people subscribe to a view and get a notification when issues enter it, including to a Slack channel (https://linear.app/docs/custom-views). Track saved views are shared with the team (components/track/SavedViewsMenu.tsx) but nobody is told when something new matches. A weekly bug triage or a Fridays-due view only works if someone remembers to open it. First scope: a Notify me toggle per view that adds a daily email or Slack summary of newly matching issues, claimed per row by the existing cron.
- 0
Track: Create projects from a project template
Linear lets a workspace save project templates with a description, milestones and a set of issues, then start new projects from them (https://linear.app/docs/project-templates). Track has issue templates per project (lib/track-templates.ts) but every new project starts empty. Teams that run the same kind of project, such as a launch or a client onboarding, rebuild its milestones and first issues each time. First scope: save a project as a template with its milestones and open issues, and a Use template choice in the new project form.
Track: Add sub-teams that roll up under a parent team
Linear supports multi-level sub-teams so teams mirror the org structure, with issues and views rolling up to the parent (https://linear-web.pages.dev/changelog/2026-04-09-multi-level-sub-teams). Track has a flat list of teams that hold the key, statuses and cycles (lib/track-data.ts), with no parent. Teams of 10 to 20 people start splitting into groups such as web and mobile and want one view of both. First scope: an optional parent team, and a team issues view that includes the issues of its children.
Track: Add team documents pinned to each project page
Linear lets a team pin shared documents to a team page, next to its issues (https://linear-web.pages.dev/changelog/2026-06-04-team-documents). Track projects have a description and updates but no place for specs, runbooks or onboarding pages, and the link from Track to Docs is only per issue (app/[locale]/app/track). A small team ends up with a doc per project that nobody finds. First scope: pin Docs pages to a project from its Overview tab, listed with title and last edit, opening in Docs.
Track: Notify and create issues from Microsoft Teams channels
Linear for Microsoft Teams lets people mention @Linear in a channel to create work and get issue notifications (https://linear-web.pages.dev/changelog/2026-04-16-linear-for-microsoft-teams). Track only reaches Teams through Flow, and the Slack integration (lib/track-slack.ts) has no Teams equivalent. Teams are common in small companies that already pay for Microsoft 365. First scope: an incoming webhook per project for activity notices to a Teams channel, then a bot command to create an issue from a message.
Track: Create a Slack channel automatically for each new project
Linear can create and link a Slack channel when a project is created, and posts project updates there (https://linear-web.pages.dev/changelog/2026-05-21-project-slack-channels). Track lets a project post issue activity to a channel chosen by hand in lib/track-slack.ts, but it never creates the channel. A small team loses time setting up one channel per project and wiring the notifications. First scope: a project setting that creates a channel named after the project through the existing Slack connection, links it and posts the project updates there.
Cal: Add a custom confirmation message with links to the booking confirmation
Calendly lets the host add instructions to the confirmation page and the confirmation email (https://calendly.com/help/how-to-customize-your-confirmation-page). We have welcomeMessage for the booking form and a redirect URL, but nothing that appears in the confirmation screen or the email, so a host cannot say 'bring your ID' or 'here is the prep doc' without a redirect. First scope: a short message per event type with plain links, shown after booking, in the email and in the manage page.
Cal: Duplicate an event type with its schedule, questions and workflow steps
Calendly has Clone on every event type in the list (https://calendly.com/help/event-types-overview), which is how people create the 15, 30 and 60 minute variants. In components/calendar/EventTypesView.tsx an event type can be opened, renamed and deleted, but the only way to get a variant is to fill the editor again, including questions and locations. First scope: a Duplicate action that copies the rules, fields, locations, schedule and workflow steps with a new slug and a Copy suffix in the title, and opens the editor.
Cal: Limit each invitee to one upcoming booking per event type
Calendly and Cal.com can limit how many future bookings the same email can hold for an event type (https://calendly.com/help/how-to-set-meeting-limits). Our meeting limits in lib/calendar-limits.ts count bookings per host and per period but not per invitee, so one person can hold every Tuesday slot. Free consultations and office hours need this to stay fair. First scope: a maxUpcomingPerInvitee rule checked at booking time, with a clear message that links to the existing booking's manage page.
Cal: Add a fixed booking date range to an event type
Calendly offers a rolling window or a fixed from/to range in Availability (https://calendly.com/help/how-to-fine-tune-your-availability-settings). Today lib/calendar.ts only supports maxFutureDays counted from today, so a workshop week or a launch campaign cannot open on a given date and close on another. This matters for recruiting weeks, office hours during an event and seasonal offers. First scope: optional startDate and endDate in the booking rules, applied in the slots engine and shown in the editor.
Cal: Show a cancellation policy on the booking page and in the confirmation email
Calendly lets you write a cancellation policy that invitees see before and after booking (https://calendly.com/help/how-to-add-a-cancellation-policy). We have a welcome message on the event type (components/calendar/PublicBookingPage.tsx) but no field that travels to the confirmation email and the manage page. Invitees at small teams often ask the same question about late changes. First scope: a policy text per event type in the three languages of the host's choice, shown on the booking form, the confirmation and the manage page.
- 0
Cal: Close cancel and reschedule from the invitee link N hours before the start
Calendly only shows a cancellation policy to invitees and does not enforce it (https://calendly.com/help/how-to-add-a-cancellation-policy), and Cal.com added a minimum notice to reschedule. Before this, the manage page in components/calendar/ManageBooking.tsx let a guest cancel or move a booking minutes before it started. A small team that blocks the calendar for a meeting needs a way to say changes close 24 hours before. First scope: a changeCutoffHours rule per event type, enforced in the manage route, with the buttons replaced by a notice. The host can still change the booking from the app.
Link: Add a count endpoint to the v1 API with the same filters as the list
Dub has GET /links/count that accepts the list filters and returns only the total (https://dub.co/docs/api-reference/endpoint/retrieve-links-count). Our API only returns total inside the paginated list, so a dashboard that wants a number must fetch a page of links too. First scope: GET /api/v1/links/count reusing parseLinkListQuery in lib/link-api-query.ts, links:read scope, grouped by an optional groupBy of tag or folder.
Link: Add bulk create, update and delete endpoints to the v1 API
Dub exposes POST, PATCH and DELETE on /links/bulk with up to 100 links per request (https://dub.co/docs/concepts/links/bulk-operations). We have bulk create and bulk actions in the app (app/api/links/bulk/route.ts) but they use the session, and the public v1 API creates one link per call. A team migrating thousands of links or tagging a campaign from a script hits the rate limit quickly. First scope: bulk create and bulk delete under links:write, up to 100 per request, with a per-item result so one bad row does not fail the batch.
Link: Add an upsert endpoint to the v1 API that creates or updates a link in one call
Dub offers PUT /links/upsert, which creates the link or updates the one with the same URL (https://dub.co/docs/api-reference/endpoint/upsert-a-link). Today an integration must GET by externalId, then POST or PATCH (app/api/v1/links/route.ts and lib/routes/v1/links/[id]/route.ts), with a race between the two calls. For a small team syncing links from a CMS or a database, one idempotent call is simpler and safer. First scope: PUT /api/v1/links/upsert keyed on externalId, falling back to destination URL, returning 200 when updated and 201 when created.
- 0
Link: Filter the v1 API links list by search, tag, folder, archived state and sort
Dub's list endpoint filters by search, tag, domain, folder and archived state (https://dub.co/docs/api-reference/endpoint/retrieve-a-list-of-links). Our GET /api/v1/links in app/api/v1/links/route.ts only accepted page, pageSize and creator, so an integration had to page through everything to find one link. First scope: q, tag, folder, archived and sort query params on the same endpoint, documented in the API reference.
- 0
Groups: Phone number masking for members
Periskope hides customer phone numbers from some roles. First scope: a workspace setting that shows only the last 4 digits to members in the inbox, lists, exports and API. Priority: low. Effort: S.
- 0
Groups: Saved chat lists as broadcast audiences
Periskope saves filtered chat lists and uses them as broadcast audiences. First scope: save the current inbox filters as a named list in the sidebar, and pick a list as recipients in a new broadcast. Priority: low. Effort: S.
- 0
Groups: More webhook events and group endpoints in the public API
Groups emits message received and ticket created or updated, and the API covers chats, messages and tickets. Periskope has 20 webhook events and group endpoints (create, members, admins, invite). First scope: chat.created, message.ack, ticket.closed, and /api/v1/groups/groups endpoints for create and members. Priority: medium. Effort: M.
- 0
Groups: Edit, delete, forward and pin messages
Periskope edits, deletes, forwards and pins WhatsApp messages from the inbox. The WAHA and Evolution APIs support them. First scope: the message actions menu plus the provider calls. Priority: low. Effort: M.
- 0
Groups: AI flags only real questions as unanswered
Groups flags a chat by time without reply. Periskope uses AI to tell questions and requests from comments like "thanks", so fewer chats are flagged. First scope: a cheap classifier on each inbound message (lib/ai.ts, off when AI is off) that skips awaiting for non-questions. Priority: medium. Effort: M.
- 0
Groups: @mentions in internal notes
Notes in a chat are visible to the team but do not notify anyone. First scope: @person in a note sends an in-app notification (the bell) and an email per the person's Track notification settings. Priority: medium. Effort: S.
- 0
Groups: Full group settings (photo, admin-only messages, approval to join)
Groups changes name and description. Periskope also sets the photo, who can send messages, who can edit info, join approval and disappearing messages. WAHA and Evolution API expose these calls. First scope: the toggles in the group page, in bulk from the list. Priority: medium. Effort: M.
- 0
Groups: Group templates with preset members and settings
Periskope creates groups from templates with the account team already inside. First scope: save a template (name pattern, members, admins, description, labels) and use it in New group and the CSV import. Priority: medium. Effort: S.
- 0
Groups: Recurring scheduled messages
Periskope repeats scheduled messages daily, weekly or monthly with an end date. Groups schedules one-off messages and broadcasts. First scope: a repeat rule on a scheduled message that creates the next send after each run (cron claim per row). Priority: medium. Effort: S.
- 0
Groups: Members see only the chats with their labels
Periskope limits what a Member sees by label, which is how teams keep private chats per account owner. First scope: a label list per person in Groups settings; members only list, search and open chats with one of their labels; admins see everything. Priority: high. Effort: M.
- 0
Groups: Round-robin assignment of new chats
Periskope assigns new chats by rules (round-robin among online agents). Today Groups assigns by hand. First scope: an on/off rule per number that rotates new chats among chosen people, skipping people marked away. Priority: high. Effort: M.
- 0
Groups: WhatsApp team inbox with group management, tickets and broadcasts
Periskope (https://periskope.app) gives B2B teams one inbox for every WhatsApp customer group. Build it in the suite: numbers by the official Cloud API or a WAHA / Evolution API gateway, shared inbox for chats and groups, bulk group management, tickets by emoji, SLAs, broadcasts, analytics and the group on its CRM record.
- 0
CRM: Import a whole HubSpot or Pipedrive account by API, with a preview and safe re-runs
Teams leaving HubSpot or Pipedrive need owners, custom properties, pipelines, companies, contacts, deals, associations and activities to come over in one go, not CSV by CSV. Paste a private app or API token, review counts and owner matching, then import in resumable chunks. Re-running updates instead of duplicating.
Other: Run PRAGMA optimize and add indexes for the remaining sort orders
The database has no sqlite_stat1, so the planner picks odd indexes: sorting contacts by name uses idx_crm_contacts_score, and the link count uses idx_links_user_creator. Several sorts build a temp B-tree: contacts by name and email with COLLATE NOCASE (lib/crm.ts:575-578), links by clicks (lib/links.ts:204), and Docs by position. That is cheap today and will not be at 50k rows. First scope: run PRAGMA optimize from the hourly cron, and add indexes on (user_id, first_name COLLATE NOCASE, last_name COLLATE NOCASE) and (user_id, clicks DESC). Priority: low. Effort: S.
Docs: Return summaries from GET /api/docs/pages and paginate
GET /api/docs/pages uses listDocPages (lib/docs.ts:250-256), which returns SELECT * with every page's content and no limit: 375 KB for 31 pages (measured). The lean listDocPageSummaries (lib/docs.ts:263) already exists. First scope: use the summaries in the list endpoint, add limit and cursor parameters, and keep content on GET /api/docs/pages/:id. Priority: low. Effort: S.
Link: Daily click rollups for analytics and record clicks after the redirect
/api/links/analytics takes 23 ms with only 6,000 clicks, 5 times slower than any other list endpoint (measured). The plan joins links and clicks and builds a temp B-tree for GROUP BY, so it grows linearly with clicks. recordClick (lib/track.ts:145) also writes synchronously before the 302, which on Turso adds a round-trip to every redirect. Dub keeps clicks in a separate events store. First scope: a link_clicks_daily(link_id, day, country, device, referrer, count) table updated in the same batch, analytics reading from it, and the clicks INSERT moved into after(). Priority: medium. Effort: M.
- 0
Agents: Index the 7-day stats queries and stop loading them in the Flow editor
listAgents (lib/agents.ts:157-176) aggregates agent_conversations with WHERE user_id AND created_at and agent_feedback with WHERE user_id, but both tables only have agent_id indexes. EXPLAIN QUERY PLAN shows a SCAN of each table, so the cost grows with every conversation. The Flow editor page (app/[locale]/app/flow/[id]/page.tsx:27-30) loads the full listAgents, stats included, just for a picker. First scope: a migration adding agent_conversations(user_id, created_at) and agent_feedback(user_id, agent_id), plus a light listAgentOptions for the pickers. Priority: medium. Effort: S.
Chat: Replace 3-second full polling with incremental fetches
ConversationClient.tsx:280-291 polls the full message list, typing status and the conversation every 3 s (about 60 requests per minute per open tab). InboxShell.tsx:139-171 polls conversations every 10 s. listChatConversations (lib/chat-data.ts:911-928) runs 2-3 correlated subqueries per row. On Turso, every open tab is constant load. Intercom pushes updates over a socket. First scope: add ?since=<updated_at> so only new messages are fetched, pause polling while document.hidden, and denormalize last_message_body onto chat_conversations. Priority: medium. Effort: M.
Board: Make Excalidraw onChange cheap and preload the editor from the list
Opening a board downloads 587 KB of gzip JS from the lazy Excalidraw chunks. BoardEditor.tsx:403-422 computes signature() over every element and calls setState('dirty') on every Excalidraw change, including appState-only changes such as picking a tool. Pressing 'r' measured an INP of 136-320 ms at 4x CPU, the only page over 200 ms. The page also calls getBoard twice (generateMetadata and the page). First scope: skip appState-only changes, use getSceneVersion for the signature, skip setState when already dirty, import() ExcalidrawCanvas on hover in BoardList, and wrap getBoard in cache(). Priority: medium. Effort: M.
Other: Avoid the full router.refresh() when the timezone cookie is first set
On the first visit to the app, components/app/TimezoneCookie.tsx:46 sets the tz cookie and calls router.refresh(). That refetches the whole RSC tree (about 159 KB, mostly i18n messages) and costs one extra server render. Measured: 724 KB total on a first visit versus 561 KB on a returning visit, with the same FCP and LCP. First scope: refresh only on pages whose dates depend on 'today', or render those dates on the client. Once i18n messages are trimmed per route, the refresh gets cheaper by itself. Priority: medium. Effort: S.
Notes: Use server data as initial state and stop loading 500 full notes
Notes scores CLS 0.133 (measured, 4x CPU): NotesApp.tsx:178,201 refetches /api/notes on mount and the main panel re-lays out. listNotes (lib/notes.ts:282-284) does SELECT n.* (full blocks and body_text) for up to 500 rows to show a snippet, and the list is not virtualized. It shares the Docs editor, so it also inherits the empty-blocks-on-SSR issue. First scope: initialize state from the server component, select only the needed columns plus substr(body_text,1,200), and paginate or virtualize the list. Priority: medium. Effort: M.
Docs: Server-render block content so the editor does not shift on load
DocsBlockEditor.tsx:97-107 writes el.innerHTML = block.content in a useLayoutEffect, so the server HTML has empty contentEditable blocks and the text only appears after hydration. Measured on a 60-block page with 4x CPU and 4G: LCP 1.85 s (other pages are around 0.9 s) and CLS 0.215. Five blocks grow from 34 to 90 px at about 2 s. The Docs tree that is portaled into #side-branch adds 0.048 more. Notion opens pages with their text already painted. First scope: render the first pass with dangerouslySetInnerHTML (keeping the anti-clobber effect for later changes) and reserve the height of #side-branch. Priority: high. Effort: S.
Other: Split app CSS per app instead of importing every app's stylesheet in the shell
app/[locale]/app/layout.tsx:12-27 imports the CSS of all 11 apps plus @xyflow/react/dist/style.css, and the root layout adds board-public.css to every page. Every app page loads 15 render-blocking stylesheets (96 KB gzip, measured). First scope: move each import to app/[locale]/app/<app>/layout.tsx, keep only app.css and color.css in the shell, and load the xyflow CSS only in the Flow editor. Priority: high. Effort: S.
Other: Dedupe the session per request with React cache() and parallelize layout reads
getCurrentUser (lib/session.ts:54-71) runs 3 times per request: in Header, in the app layout and in the page. Each call does getServerSession, getUserByEmail and resolveWorkspace (up to 4 queries), and in DEMO_MODE it also writes UPDATE users SET seeded_at. The app layout adds countDueTasks and about 12 sequential getTranslations calls. Locally TTFB is 20-40 ms, but on Turso every query is a round-trip. First scope: wrap getCurrentUser in cache(), use Promise.all in app/[locale]/app/layout.tsx, and keep the seeded flag in memory for the demo user. Priority: high. Effort: S.
Other: Send only the route's i18n namespaces to the client
i18n/request.ts loads all 22 namespaces (568 KB of JSON in es) and app/[locale]/layout.tsx:75,95 passes them whole to NextIntlClientProvider. Every HTML document weighs 160-200 KB gzip, and about 158 KB gzip of that is messages, on public pages too (/@handle, booking, help center). Measured in docs/benchmark/performance.md (4x CPU, 4G). First scope: the root layout picks only 'common', each app layout adds a nested provider with its own namespace, and public pages add theirs. Server components keep reading everything through getTranslations. Expected: -130 to 150 KB gzip per document, and less on every router.refresh. Priority: high. Effort: M.
- 0
Track: Auto-close stale issues and auto-archive completed ones
Linear lets each team auto-close issues untouched for N months and auto-archive completed issues after a period, which keeps lists short without manual cleanup (https://linear.app/docs). Track keeps every issue in its list forever unless deleted. First scope: per-project settings for auto-close (move to a cancelled status after 1, 3 or 6 months without activity, with a comment explaining why) and auto-archive (hide done/cancelled issues older than N months from default views, still searchable and restorable), run from the hourly cron with compare-and-set claims. Priority: low. Effort: S. Compared with Linear.
- 0
Track: Set SLAs by priority and flag issues about to breach
Linear supports issue SLAs and triage responsibility so urgent bugs get a response time (https://linear.app/pricing, Triage section); Jira Service Management triggers automations on SLA breach. Track has no response or resolution target. First scope: per-project SLA per priority (e.g. urgent 24 h, high 3 days) counted from creation or triage acceptance, a countdown chip on rows and the issue page, an "SLA" filter and sort, a notification to the assignee and project owner at 80 % and at breach, and an issue.sla_breached event for Flow. Priority: medium. Effort: S. Compared with Linear, Jira.
- 0
Track: Link CRM companies and contacts to issues as customer requests
Linear Customer Requests link feedback to issues and projects, show a customer page with every request, and let views sort issues by customer count, tier or revenue (https://linear.app/docs/customer-requests). zerosoft already has the CRM and Chat, so this can work without any third-party sync. First scope: a "Requested by" field on issues that links CRM companies or contacts (and the Chat conversation when created from Chat), a request count and summed deal value column in lists with sort, and a Track section on the CRM company page listing its requested issues and their status. Priority: medium. Effort: M. Compared with Linear.
- 0
Track: Suggest labels, assignee and possible duplicates for issues in triage
Linear Triage Intelligence proposes labels, assignee, project and duplicates for incoming issues (https://linear.app/pricing). Track's triage queue (issues from Chat, Flow, API, GitHub, Agents) is fully manual. First scope, only when AI is enabled for the workspace: on each triage issue, show suggested labels and assignee based on the last 200 accepted issues of the project, and the top 3 similar open issues by title and description with a one-click "Mark duplicate"; accept applies the suggestions in one click. Log which suggestions were accepted. Priority: medium. Effort: M. Compared with Linear.
- 0
Track: Add private projects and guest members limited to the projects they are invited to
Linear Business has private teams and guest accounts (https://linear.app/pricing); GitHub Projects and Jira have per-project permissions. In Track every workspace member sees and edits every project (canAccessTrackProject checks only the workspace), so teams that work with contractors or clients cannot use it. First scope: a project visibility setting (workspace / private with a member list), guest role that only sees projects they are added to and cannot see Cmd+K results or Inbox items from others, and enforcement in every /api/track route plus search and notifications. Priority: medium. Effort: M. Compared with Linear, Jira.
- 0
Track: Show cycle velocity, burn-up and cycle time so teams can plan the next cycle
Linear Insights charts issue count and effort, cycle time, lead time, triage time and issue age, with slicing, segments and a burn-up, and exports to CSV (https://linear.app/docs/insights); cycles show scope and completion over time. Track's overview only shows done/total for the current cycle. First scope: on each cycle, a burn-up of scope vs completed points or issues (from the activity log), velocity of the last 3 completed cycles shown when creating the next one, and a cycle-time / lead-time panel per project with median and p75, using status category changes already logged in track activities. Priority: high. Effort: M. Compared with Linear, Jira.
- 0
Track: Add projects with lead, target date, health and written updates, plus a roadmap across them
In Linear a project has a lead, members, start and target dates, status, and periodic updates marked On track / At risk / Off track with reminders to the lead (https://linear.app/docs/initiative-and-project-updates); initiatives group projects into a roadmap. A Track project is closer to a Linear team: it has milestones but no lead, target date, health or updates, and there is no view across projects. First scope: lead, target date and status on track_projects; a project update composer with health and auto-included progress since the last update; a weekly reminder to the lead; a roadmap page with one bar per project (start to target) and its latest health. Priority: high. Effort: L. Compared with Linear.
- 0
Track: Create issues from Slack messages and post issue activity to a Slack channel
Linear's Slack integration creates an issue from any message (with a synced comment thread), posts team and project activity to a channel, sends personal notifications as DMs and unfurls issue links (https://linear.app/docs/slack). Track only reaches Slack if someone builds a Flow with the generic Slack node. First scope: one-click Slack connect in Track › Settings, a message shortcut "Create Track issue" that opens a modal with project, priority and assignee and links back to the message, per-project channel notifications for created / completed / commented, and unfurls for issue URLs. Priority: high. Effort: M. Compared with Linear.
- 0
Track: Update lists, board and issue pages live when a teammate changes something
Linear syncs every change to every open client instantly. In Track the project list, board and issue page only change on reload: in a browser test, renaming an issue from a second session left the open list showing the old title after 15 seconds and a focus event (components/track/ProjectPageClient.tsx has no polling or subscription). Two people triaging the same board see different states and overwrite each other. First scope: a per-workspace change feed (SSE, or polling every few seconds while the tab is visible) carrying issue id + updated_at, merged into useIssueStore; refetch on window focus; on the issue page, show new comments and a notice when the description changed underneath an open editor. Priority: high. Effort: M. Compared with Linear.
- 0
Track: Import issues from Linear, Jira, GitHub Issues and CSV, and export a project to CSV
Linear has guided importers for Jira, GitHub Issues, Asana, Shortcut and Linear itself that bring issues, projects, labels and users, with user mapping and an undo window (https://linear.app/docs/import-issues). Track has no importer and no export: a team switching over has to retype its backlog, which is the first reason not to switch. First scope: a CSV importer in Track › Settings with column mapping (title, description, status, priority, assignee by email, labels, due date, estimate, parent) into one project, plus a Linear importer using a personal API key (issues, comments, labels, states mapped by category) and a GitHub Issues importer per repo. Export: CSV of the current list view. Priority: high. Effort: M. Compared with Linear, Jira.
Robots: Expose robots as tools over MCP
Apify runs a hosted MCP server at mcp.apify.com that lets Claude, Cursor or ChatGPT search, run and read Actors (https://docs.apify.com/integrations/mcp). Robots can already be run by Agents inside zerosoft and through /api/v1/robots, but external AI clients cannot use them. First scope: an MCP endpoint that lists the workspace's robots as tools with their inputs as the schema, runs them through the existing queue and returns the output, behind an API key with the robots:run scope. Priority: low. Effort: S. Compared with Apify.
Robots: Allow longer runs and more pages for catalog crawls
Browse AI allows tasks up to 60 minutes on every paid plan (https://www.browse.ai/pricing); Apify runs are bounded only by the plan's compute. Robots caps a run at 180 seconds, 20 pages and 1,000 list rows (lib/robots-schema.ts), so a 30-page catalog or a detail-page crawl over 100 items cannot finish in one run. First scope: a long-run mode (up to 15 minutes, 200 pages, 10,000 rows) available to scheduled and API runs, with progress saved per page so a timeout keeps what was extracted. Priority: medium. Effort: M. Compared with Browse AI / Apify.
- 0
Robots: Paginate with Load more buttons and infinite scroll in Extract list
Browse AI paginates numbered pages, 'load more' buttons and infinite scroll (https://www.browse.ai/features). Extract list and Extract table only follow a Next button (lib/robots-schema.ts pagination), so feeds and modern catalogs need a hand-built loop of scroll and click steps that does not dedupe rows. First scope: a pagination mode select (Next, Load more, Scroll) where Load more clicks until the button is gone and Scroll scrolls until no new cards appear, both deduping rows by card text. Priority: medium. Effort: S. Compared with Browse AI.
- 0
Robots: Bulk run from a CSV upload in the UI and raise the batch limit
Browse AI bulk runs take thousands of inputs uploaded from a sheet or CSV (https://www.browse.ai/docs/api/v2). Robots bulk runs exist only in the v1 API and cap at 100 inputs per batch (MAX_BULK_INPUTS in lib/robots-schema.ts). Running a robot over a list of 1,000 product URLs is the common case. First scope: a 'Run for many' dialog that accepts a CSV, maps columns to inputs, shows batch progress and lets you download all results as one CSV, with the batch limit raised to 1,000 and runs queued. Priority: medium. Effort: S. Compared with Browse AI.
- 0
Robots: Keep an accumulated table of rows across runs with a dedupe key
Browse AI stores every run's rows in a per-robot Table with tabs, filters and export (https://help.browse.ai/en/articles/10477537-how-to-view-manage-and-export-your-scraped-data); Apify appends to named datasets (https://docs.apify.com/platform/storage/dataset). In Robots each run keeps its own output and CSV, so a price history or a growing lead list has to be assembled by hand. First scope: a Data tab that merges a list variable across runs by a chosen key column, with first seen, last seen and changed columns, filters and CSV export. Priority: medium. Effort: M. Compared with Browse AI / Apify.
- 0
Robots: Email a highlighted diff when a monitored robot's output changes
Browse AI monitors send a color-coded email with added, changed and removed rows (https://help.browse.ai/en/articles/10524268-how-to-set-up-monitoring-notifications). Robots already computes the diff and emits robot.output.changed (lib/robots-diff.ts), but nobody hears about it unless they build a flow on that event. First scope: an 'Email me when the data changes' checkbox next to the schedule that sends the diff (before and after per key, rows added and removed) to the robot owner, throttled to one email per run. Priority: medium. Effort: S. Compared with Browse AI.
- 0
Robots: Send extracted rows to Google Sheets or Airtable from the robot
Browse AI keeps Google Sheets and Airtable in live sync with a robot's table (https://www.browse.ai/pricing). In Robots the only path is building a flow from the 'Robot data to Google Sheets' template, which needs a service account and only appends. First scope: a Destination section in the robot's Advanced tab that picks a Sheets credential and range and appends (or replaces) the rows of a chosen list variable after each successful run, reusing the Flow Sheets client. Priority: medium. Effort: S. Compared with Browse AI.
Robots: Route runs through residential proxies by country and handle captchas
Browse AI uses rotating residential proxies with location targeting and solves supported captchas and Cloudflare challenges (https://help.browse.ai/en/articles/10452902-what-websites-work-with-browse-ai); Apify sells datacenter, residential and SERP proxies with country targeting (https://docs.apify.com/proxy). Robots runs from the runner's own IP (lib/robots-runner.ts), so marketplaces and real estate portals block it or show the wrong country. First scope: a proxy credential type (URL, user, password) and a per-robot proxy and country setting passed to Chromium launch, then a captcha-solver credential used when a challenge is detected. Priority: high. Effort: M. Compared with Browse AI / Apify.
- 0
Robots: Generate a robot from a URL and a description of the columns
Browse AI Table Studio takes a URL, proposes tables and columns with sample rows and lets you approve each column before saving (https://help.browse.ai/en/articles/10486935-how-browse-ai-uses-ai). Robots has no AI in the builder; the backlog idea for natural-language act and extract steps is about runtime steps. First scope: a 'Start from a URL' dialog that loads the page in the runner, sends a trimmed DOM to the AI gateway and returns goto + extract_list (with fields and Next pagination) as unsaved steps the user can run and edit. Priority: high. Effort: M. Compared with Browse AI.
- 0
Robots: Pick what to extract by clicking on a live preview of the page
Browse AI builds robots by pointing at the page: Capture List and Capture Text in its Chrome extension (https://help.browse.ai/en/articles/10492719-how-to-build-a-robot-using-the-chrome-extension). Robots asks for CSS selectors typed by hand, and the bookmarklet recorder only captures clicks, fills, selects and keys, not extractions (lib/robots-recorder.ts). A user who does not know CSS cannot build the first robot. First scope: a 'Pick on page' button in Extract and Extract list that opens a sandbox screenshot with element boxes; clicking a card proposes the card selector and its fields with generateSelector, then the same mode inside the bookmarklet. Priority: high. Effort: L. Compared with Browse AI.
- 0
Notes: Create a note linked to a Cal booking or event
Evernote's calendar pop-up creates a note linked to a Google or Outlook event (December 2025, https://evernote.com/blog/2025-recap). Our Cal app has bookings with attendees, but there is no way to open a meeting note from them. First scope: a 'Meeting note' action on a Cal booking that creates a note titled with the event and date, with attendees and a link back, and lists linked notes on the booking. Priority: low. Effort: S. Compared with Evernote.
- 0
Notes: Record audio in a note and transcribe it
Evernote records audio with pause/resume and transcribes it with speaker recognition (https://evernote.com/blog/2025-recap, https://releasebot.io/updates/evernote); Apple Notes records and transcribes too (https://support.apple.com/guide/notes/record-and-transcribe-audio-apdb5106e334/mac). Notes has file blocks (10 MB cap) but no recorder and no transcription. First scope: a 'Record' block that uses MediaRecorder in the browser, uploads the audio as an attachment, and, when AI is set up, adds a transcript below it that is searchable and used by Ask your notes. Priority: low. Effort: M. Compared with Evernote.
- 0
Notes: Select several notes to move, tag, pin or trash them at once
Evernote added multi-select bulk actions on mobile in 2025 (https://evernote.com/blog/2025-recap) and has them on desktop. The Notes list only acts on one note at a time (row menu in components/notes/RowMenu.tsx), which makes cleaning up an ENEX import slow. First scope: Shift/Cmd-click and checkbox selection in the list, a bar with Move to notebook, Add tag, Pin, Trash and Export, keyboard support (Shift+Arrow to extend, Esc to clear) and one undo toast for the batch. Priority: medium. Effort: S. Compared with Evernote.
- 0
Notes: Create notes from templates
Evernote offers templates when you create a note (meeting, journal, trip) and made template creation easier in 11.31 (https://releasebot.io/updates/evernote). Docs already has built-in and workspace templates (lib/docs-templates.ts), but Notes always starts empty. First scope: a template picker on the 'New note' split button with a few personal templates (meeting, daily log, reading notes), 'Save as template' from any note, and templates that preset tags and notebook. Priority: medium. Effort: S. Compared with Evernote.
- 0
Notes: Group notebooks into stacks
Evernote groups notebooks into stacks, and an ENEX import of a long-time user brings dozens of notebooks. Notes has a flat notebook list in the sidebar (components/notes/NotesApp.tsx), so after importing 30 notebooks the sidebar is a long scroll. First scope: a stack (or one level of parent notebook) that notebooks can be dragged into, collapsible in the sidebar, with 'notebook:' search matching a stack name to include all its notebooks, and keep the stack name when importing ENEX files that carry it. Priority: medium. Effort: S. Compared with Evernote.
- 0
Notes: Export a notebook or all notes as a Markdown or ENEX zip
Evernote exports notebooks as ENEX or HTML from desktop, Apple Notes exports PDF and Markdown, and Bear exports many formats (https://bear.app/faq/export-your-notes/). Notes exports one note at a time to Markdown (app/api/notes/[id]/export). People moving years of notes in want to know they can get them out the same way. First scope: 'Export notebook' in the notebook menu and 'Export all notes' in Notes settings, returning a zip with one .md per note (front matter with tags, created, updated, source_url), attachments in a folder, and an optional .enex file so the export round-trips into Evernote. Priority: medium. Effort: S. Compared with Evernote.
- 0
Notes: Save searches with operators as smart lists in the sidebar
Evernote has saved searches and Apple Notes has Smart Folders filtered by tags, checklists and dates (https://support.apple.com/guide/notes/use-smart-folders-apd58edc7964/mac). Notes already parses tag:, notebook:, intitle:, created:, updated: and todo: (lib/notes-query.ts) but a query is lost when you clear the box. First scope: a 'Save search' action next to the search field that stores name + query per user, lists saved searches in the sidebar under Tags, keeps them in sync with renames of tags and notebooks, and opens them with Cmd+K. Priority: medium. Effort: S. Compared with Evernote.
Notes: Search text inside images and PDF attachments
Evernote searches text inside images and PDFs and, since 11.32, auto-generated image captions (https://releasebot.io/updates/evernote); Bear Pro does OCR on images and PDFs (https://bear.app/faq/features-and-price-of-bear-pro/). Notes stores attachments (app/api/notes/[id]/files) but search only looks at title, text blocks and tags, so receipts and scans imported from ENEX never show up. First scope: on upload (and for ENEX resources on import), extract PDF text and run OCR on images in a background job, store the text per attachment, include it in notes search and Ask your notes, and show 'Found in attachment' in the result. Priority: high. Effort: M. Compared with Evernote.
- 0
Notes: Add due dates to checklist items and a Tasks view across notes
Evernote tasks live inside notes with due dates, recurrence, reminders and flags, and a central Tasks view lists them all (https://help.evernote.com/hc/articles/1500003792141). In Notes a checklist item is a to-do block with no date, and reminders are one per note (lib/notes-reminders.ts). 'Send to Track' covers team work, but personal tasks have nowhere to live. First scope: a due date on to-do blocks (typed with a date picker or '@' shortcut), a 'Tasks' entry in the Notes sidebar listing open items from every note grouped by Overdue / Today / Upcoming / No date with a link back to the note, and reuse the reminder cron to notify on the due time. Priority: high. Effort: M. Compared with Evernote.
Link: One-click DNS setup for custom domains
Dub added Domain Connect for one-click DNS on Vercel and Cloudflare in August 2026 (https://dub.co/changelog/domain-connect). Our domain setup shows records to copy and a verify button. Scope: detect the DNS provider, offer Domain Connect or the Cloudflare API with a token, then trigger verification automatically. Priority: low. Effort: M. Compared with Dub.
- 0
Link: Show a live stream of clicks as they happen
Dub analytics include a real-time event stream (https://dub.co/links). Our click log is paginated and only updates on reload. Scope: a live mode on the click log and workspace analytics that polls or uses server-sent events every few seconds, highlighting new rows with country, referrer and device. Priority: low. Effort: S. Compared with Dub.
- 0
Link: Keep an activity log of link changes per workspace
Dub includes audit logs on Enterprise (https://dub.co/pricing). We record who created a link but not who changed its destination, slug or targeting later, which matters when the link is printed on a QR. Scope: write an entry on create, update (field diff), archive and delete, show a History tab in link detail and a workspace-wide log filterable by member. Priority: low. Effort: M. Compared with Dub.