Skip to content
Documentation
MCP Server

GitDailies MCP Server

The GitDailies MCP Server lets AI assistants and agents, such as Claude, ChatGPT, Cursor, and GitHub Copilot, query your GitDailies engineering metrics and your own GitHub work directly. It uses the Model Context Protocol, the open standard for connecting AI tools to data sources.

Once it’s connected, you can ask your assistant questions in plain language:

  • “What’s waiting for my review, oldest first?”
  • “How has our PR pickup time changed by week over the last quarter?”
  • “Why did lead time for change spike in August? Show me the PRs behind it.”
  • “Summarize what I merged and reviewed this week for my standup.”

Your assistant picks the right GitDailies tools, fetches the data, and answers using the same metrics you see in the GitDailies web app.

Connect your AI assistant
Add the GitDailies MCP Server to Claude, ChatGPT, Cursor, VS Code, or any MCP client.
Tools reference
The metrics, drill-down, and personal tools that the server exposes to your agent.
Example prompts
Questions to try, from your daily standup to quarterly DORA reviews.

What your agent can access

  • Team metrics: PR trends, review trends, open PR status, and DORA metrics, with the same filters and aggregations as the Metrics API.
  • Drill-downs: the individual PRs, reviews, deploys, or incidents behind any data point, so your agent can explain why a metric moved, not just that it moved.
  • Your own work: your open PRs, your outstanding review requests, what you’ve merged and reviewed recently, and your triggered personal alerts.

The server is read-only. It can’t change anything in GitHub or GitDailies.

How is this different from GitHub’s MCP server?

GitHub’s MCP server is built for working with code: reading files, opening issues, and managing PRs. It gives your agent raw data, one API call at a time. To answer a question like “how has our review time changed this quarter?”, an agent would have to page through thousands of PRs, reviews, and timeline events, then join them and do the math itself. That fills its context window, runs into rate limits, and risks hallucinated numbers.

The GitDailies MCP server gives your agent metrics that are already computed:

  • Accurate and consistent: metrics are calculated by tested, deterministic code, and match your GitDailies dashboards and reports.
  • Efficient: one call returns a whole trend, filtered and aggregated by team, repo, user, label, branch, and time period.
  • Traceable: drill-down tools return the exact PRs, reviews, and deploys behind any number.
  • Beyond GitHub: DORA metrics, using deploy and incident events, and your triggered GitDailies alerts.

Your agent can then focus on what it’s best at: spotting patterns and explaining them. The two servers work well together. Use GitHub’s MCP server for working with code, and GitDailies for metrics.

Security and access

  • You sign in with OAuth when you first connect. There are no API keys to create, copy, or rotate.
  • Each connection is bound to one GitDailies install (one GitHub organization). Your agent only sees repos that the install follows in GitDailies.
  • Personal tools (“my PRs”, “my reviews”, “my alerts”) describe the GitHub user who authorized the connection.
  • MCP access is included in the Max plan. If the server reports that MCP access isn’t included, an organization admin can upgrade from the Settings page of the GitDailies web app. You can also reach out to support@gitdailies.com for a free trial.

Need help?

Contact us at support@gitdailies.com with any questions.