Connect Bitbucket with Unblocked to give AI agents codebase context

Bitbucket contributes two different kinds of evidence to an engineering answer: repository state and the pull-request conversations behind that state. Unblocked connects them so agents can move from an implementation detail to the review history that explains it.

Bitbucket context doesn’t live alone. Unblocked links pull requests and their discussions to the code they touch, then combines them with your docs, tickets, and conversations from every other connected source. Instead of raw search results, agents and developers get one reconciled answer with Bitbucket sources cited.

Once connected, any MCP-compatible coding agent — Claude Code, Cursor, Windsurf, GitHub Copilot, Codex — can pull this context through a single Unblocked MCP server connection.

What you get

• Repository context — Unblocked indexes your source code and keeps it up to date as you push changes. The default branch is the source of truth.
• Pull request history — Every pull request, its review comments, and merge decisions become searchable context. Agents can reference past PRs to understand why code works the way it does.
• Cloud and Data Center — Bitbucket Cloud connects via OAuth in minutes; Bitbucket Data Center is supported on the Enterprise plan with an admin install and optional service account.
• Cross-source synthesis — Bitbucket data is combined with Slack threads, Jira tickets, Confluence docs, and Sentry errors into a single knowledge graph.

Need screenshots or troubleshooting?This is a concise connection summary. Follow the complete setup guide ↗

Prerequisites

• An Unblocked account (free to start).
• Bitbucket workspace administrator access to install Unblocked. Bitbucket Data Center requires the Enterprise plan.
• An MCP-compatible coding agent (Claude Code, Cursor, Windsurf, etc.) to use the context in your editor.

Step 1: Connect Bitbucket to Unblocked

1. As a Bitbucket workspace administrator, navigate to Team Settings → Data Sources in Unblocked and select Bitbucket.
2. Authorize the Unblocked app via OAuth (sign in to Bitbucket if prompted). This lets Unblocked read your workspaces and repositories but doesn’t index anything yet.
3. Select the workspace you want Unblocked to use.
4. Choose the repositories to index — Unblocked preselects repos active in the last 6 months; forks and archived repos are excluded — then click Save and Continue.

Step 2: Install the Unblocked MCP server

On macOS, one command auto-detects and configures your supported agents (Claude Code, Cursor, Windsurf, GitHub Copilot, Codex, VS Code):

The install script supports macOS only. On Windows or Linux, add the server manually using the configuration below, or follow the manual install guide ↗.

claude mcp add -s user --transport http unblocked https://getunblocked.com/api/mcpsse

{
  "mcpServers": {
    "unblocked": {
      "url": "https://getunblocked.com/api/mcpsse"
    }
  }
}

https://getunblocked.com/api/mcpsse

To give Claude Code, Cursor, GitHub Copilot, Codex, or another MCP-compatible AI coding agent access to Bitbucket context, connect Bitbucket to Unblocked and add the single Unblocked MCP server to your agent.

Unblocked acts as the MCP context layer for your Bitbucket data, searching repositories, pull requests, code history alongside your other connected sources and returning cited results through these tools:

• context_researchMulti-source research across your Bitbucket repositories, pull requests, docs, and messaging. Returns a synthesized answer with citations, so agents understand not just what the code does but why it changed.
• context_get_urlsFetches the full content behind Bitbucket pull request and file URLs your agent encounters — plus Jira, Linear, Notion, Confluence, Slack, and more.

Questions you (or your agent) can ask Unblocked about your Bitbucket data:

• Why was the caching layer rewritten, and which pull request introduced it?
• What decisions were made about the database migration approach?
• Which pull requests have touched the checkout service in the last 30 days?
• What did reviewers raise on the last PR to the auth repository?