mcp-sentry

mcp-sentry is a Model Context Protocol (MCP) server that connects to your Sentry.io account to fetch and analyze issues. It exposes tools like get_sentry_issue, which retrieves and analyzes a specific Sentry issue by ID or URL, returning details such as title, issue ID, status, level, first and last seen timestamps, event count, and a full stacktrace. It also provides get_list_issues, which fetches issues for a given project slug and organization slug, returning a list with key attributes like title, issue ID, status, level, time bounds, event count, and basic information. A dedicated sentry-issue prompt is available (sentry-issue) to retrieve and format issue details for conversational context. To operate, configure an MCP server (via uvx, Docker, or Python) with your Sentry token, project slug, and organization slug, then invoke the appropriate commands or prompts to fetch the data you need.

Prerequisites:

• Python 3.8+ (for pip installation and python -m mcp_sentry usage)
• pip or uv (optional for development)
• Access token for Sentry.io with read permissions
• Optional: Docker if you prefer containerized usage

Install via Python (recommended):

1. Install the package

   pip install mcp-sentry
   

2. Run the server directly

   python -m mcp_sentry --auth-token YOUR_SENTRY_TOKEN --project-slug YOUR_PROJECT_SLUG --organization-slug YOUR_ORGANIZATION_SLUG
   

Alternative installation (uv/uvx):

1. Install uv if you do not have it

   pip install uv
   

2. Run using uvx directly (no permanent install needed)

   uvx mcp-sentry --auth-token YOUR_SENTRY_TOKEN --project-slug YOUR_PROJECT_SLUG --organization-slug YOUR_ORGANIZATION_SLUG
   

Using Smithery (optional):

1. Install via Smithery for Claude Desktop automation

   npx -y @smithery/cli install @qianniuspace/mcp-sentry --client claude
   

Prerequisites recap: ensure you have an active Sentry token, know your project slug and organization slug, and have the runtime (Python, uv, or Docker) installed as appropriate.

Tips and common issues:

• Ensure your SENTRY_TOKEN has read access to the required projects.
• When using uvx or Docker, double-check that the command line arguments correctly map your project_slug and organization_slug to the Sentry account.
• If you encounter environment path issues in Zed, you may need to set PYTHONPATH when running via pip installation as shown in the docs.
• For debugging, you can use the MCP inspector (npx @modelcontextprotocol/inspector uvx mcp-sentry ... ) to inspect runtime behavior and responses.
• The readme shows multiple deployment options; pick the one that aligns with your workflow (local Python, uv/uvx, or Docker).
• Replace placeholders YOUR_SENTRY_TOKEN, YOUR_PROJECT_SLUG, and YOUR_ORGANIZATION_SLUG with your actual credentials before running.