Release Intelligence turns your recent Linear work into a changelog draft. When you complete a project in Linear, or click Generate Changelog, it collects every issue and project completed since your last published changelog and writes the entry.
Each completed project gets a two-paragraph summary. Completed issues are grouped by their Linear label, or by the product area the AI infers, and listed as short bullets with readable titles.
How it works
Automatic: complete a Linear project, and a changelog draft appears in Productlane. Merging a pull request does not start a draft by itself.
Manual: click Generate Changelog on the changelog screen or inside an entry. It collects everything completed since your last published changelog, or the last three months if you have none.
Generation: the AI writes the draft from the titles, descriptions, and labels of those issues and projects. With GitHub connected, it also reads the merged pull requests (up to 40) and non-merge commits (up to 50) since your last changelog. A pull request needs no linked Linear issue to be included. Commits whose message starts with "Merge pull request", "Merge branch", or "Merge remote-tracking branch" are left out.
Review and publish: edit the draft, then publish when ready. Because of the 40 pull request and 50 commit caps, the draft may not list every change, so review it before you publish.
Language
Drafts are written in your workspace's default language, set under Settings > Multilingual. Product names, brand names, and code samples stay untranslated. Without a default language, drafts are written in English.
Requirements
Linear connected under Settings > Integrations > Linear.
The Pro plan or above, or a trial.
Generation draws on your monthly AI budget. See AI Usage & Credits.
Search your changelogs
The search field at the top of the changelog sidebar searches titles and notes. Results are ranked by relevance and show the entry date, a Draft pill for unpublished entries, and the matching text. Clear the field to return to the full list.
Changelogs through MCP
An AI client connected over MCP, such as Claude or Cursor, can draft, edit, publish, and broadcast changelog entries for you. Grant the changelogs:read and changelogs:write scopes when you connect it.