Connect Linear and BeforeQuery indexes completed issues and documents — bug fixes, feature context, and specs your team already wrote — so the assistant can answer “was this fixed?” and “how does X work?” with citations back to Linear.

​
Prerequisites

  • A Linear API key (lin_api_…) — create one in Linear under Settings → Security & access → Personal API keys (or use a dedicated service account’s key). There are no scopes to choose: the key carries its creator’s own access, so make it on an account that can see the teams you want indexed.
  • A BeforeQuery workspace owner role.

​
Data ingested

Issues become one document each:
  • Issue URL, and IDENT-123: title as the title
  • Status and labels
  • Description, then every comment with its author
Documents become one each: the document’s URL, its title, and its body.

​
Index Linear issues and documents

1

Add the source

In the dashboard, open Knowledge Base → Sources → Add source and choose Linear. Paste the API key and, optionally, a list of team keys (e.g. SUP) to restrict indexing; leave empty to index all teams the key can read.
2

Sync and verify

Click Add. Completed issues and documents are normalized into markdown and embedded. Test a known fixed-bug question in the Playground.

​
Resolved support tickets via Linear

If your support flow resolves tickets in Linear, you can also connect Linear as a workspace-level helpdesk integration (Workspace → Helpdesk Integrations → Linear) and add a Resolved Tickets source that references it. That connection additionally powers copilot drafts and deflection.

​
Keeping content in sync

Linear re-syncs every 10 minutes — there is no schedule to configure, and Sync on the source’s page runs one now if you cannot wait. Only changed content is re-embedded. See Sources Overview.

​
Troubleshooting

  • Issues missing? Only completed issues are indexed — in-progress work is skipped by design. Also verify the key’s team access.
  • Internal roadmap details in answers? Mark Linear-derived documents restricted so only workspace members get those answers, or isolate the source in an internal source group.