Cost of AI per merged PR
Pull merged PRs with and without AI by team from LinearB, join them with the tool costs in my “AI spend” sheet, and chart AI cost per merged PR for each team.
Every Monday at 9:00, post each team’s delivery summary for the last week to its Slack channel. Flag pickup time over 12 hours and list PRs waiting more than two days.
get_teamsget_metricsget_metrics_benchmarkssearch_pull_requests+Payments · delivery brief for the week of Oct 12
⚠️ Pickup time is over the 12 hour target for the second week running.
Waiting more than two days:
Retry budget for ledger sync · 2d 6h
Refund webhook timeouts · 2d 2h
Sample data · LinearB tool names as the MCP server lists them
Trusted by the teams transforming software delivery
LinearB brings
PagerDuty brings
What you can build with both
Pull merged PRs with and without AI by team from LinearB, join them with the tool costs in my “AI spend” sheet, and chart AI cost per merged PR for each team.
Every weekday at 10:00, find PRs that have waited more than a day for review in LinearB, and send each reviewer a Slack message with links to theirs.
Take the PRs in our latest release from LinearB, match them to their Jira issues, and write customer-facing release notes on our Notion changelog page.
Compare Zendesk ticket volume for the last two weeks with LinearB releases over the same period, and tell me which release lines up with the jump.
Each morning, check change failure rate for every service in LinearB. If one passes 10%, use Zapier to open a Jira ticket for the team that owns it.
Build me a page I can open every morning, with delivery, incidents and AI adoption by team from LinearB, the current Jira sprint and who’s on call.
For each 1:1 on my calendar tomorrow, summarize what that person’s team shipped in LinearB this sprint, with waiting reviews and pickup time.
Write the description for this PR and link the Jira issue it closes. Then check in LinearB how long reviews usually take on this repository.
Each person connects with their own LinearB account through OAuth, so the MCP server returns only the data their LinearB role and team scope allow.
The LinearB MCP server doesn't connect to Slack, Salesforce or the other servers you add. Your AI tool calls each one separately and puts the answers together.
The numbers come from the same LinearB data behind your dashboards, so you can open the dashboard and confirm a figure before you share a report.
Claude, Claude Code, ChatGPT, Codex, Cursor, VS Code, Windsurf, Zed, Cline and any other client that supports MCP. Each person connects with their own LinearB account through OAuth. In Claude, custom connectors need a paid plan. Connect the MCP server
Your AI tool does the combining. Add the MCP servers or connectors for your other tools to the same client, and one prompt can use LinearB data alongside theirs.
Yes. The LinearB MCP server runs next to any other MCP server in your client, such as Claude Desktop or ChatGPT.
Metrics and benchmarks, pull requests, branches, releases, incidents, services, teams, contributors, issues, and project management entities such as epics and initiatives. It reads the git, project management and incident data you've connected to LinearB. Every git provider and project management tool LinearB supports works with it, including GitHub, GitLab, Bitbucket, Azure DevOps and Jira.
Yes, when people connect through OAuth. The MCP server returns only the data each person's LinearB role and team scope allow. For clients that don't support OAuth, a Company Admin can use an organization-scoped API key instead.
Save the prompt as a scheduled task or a reusable page in your AI tool, and keep the teams and timeframe explicit. The LinearB MCP server comes with ready-made prompts for release, urgent PR, lead time and merge rate reports, and the LinearB prompt library has more.
The numbers come from LinearB, but your AI tool writes the answer, so results can be incomplete in edge cases. Ask for one metric, team and timeframe at a time, and check figures in your LinearB dashboards before you make a decision on them.