GitHub logo

Know where engineering stands without leaving the conversation

Connect GitHub's MCP server to Spark and ask about pull requests, issues, and CI runs in plain language.

Overview

The GitHub MCP connector provides read-only access for getting information on specific pull requests, issues, CI runs, branch state, and more — all through natural language prompts. Whether you're an engineering manager working in Productboard or a product manager pushing production code, the connector keeps you in-the-flow with fewer side trips into GitHub.

To push features from Productboard to GitHub as issues, see the GitHub integration. To ask Spark about your product's implementation and surface technical requirements, see GitHub codebase analysis.

Ask about pull requests, issues, and CI runs

Ask questions like "What's blocking the checkout PR?" or "Did the last CI run on main pass?" and get a structured answer in chat. The connector works best on a named repository, pull request, issue, or CI run.

Routed alongside codebase analysis

Spark sends questions about object state — pull requests, issues, CI — to GitHub MCP, and questions about how the product works under the hood to codebase analysis. You ask the question, and Spark picks the source.

Setup

Any Spark user can connect GitHub. Connectors are set up per user, and no admin setup is required in Productboard.

  1. From any Spark chat, select the Integrations & Connectors icon beneath the prompt input field.
  2. Select + Add integrations & connectors and open the MCP Connectors tab.
  3. Click Connect next to GitHub and authorize with your GitHub account. You'll return to Productboard with the connection active in every Spark chat.

The connector supports GitHub.com. GitHub Enterprise and self-hosted GitHub are not currently supported.

More integrations in this category