MCP Server Archived reference

Sentry

MCP server that retrieves Sentry.io issues by ID or URL, returning details and full stacktraces to MCP clients like Claude Desktop, VS Code and Zed.

pip install mcp-server-sentry
Sentry preview image

What it is

mcp-server-sentry is a Model Context Protocol server for retrieving and analyzing issues from Sentry.io. It provides tools to inspect error reports, stacktraces, and other debugging information from your Sentry account. It exposes one tool, get_sentry_issue, and one prompt, sentry-issue. The repository is in the archived servers collection.

Who it's for

  • Developers who want to inspect Sentry issues and stacktraces from an MCP client
  • Claude Desktop, VS Code or Zed users who want Sentry debugging context in their assistant

Requirements

Requirements

  • A Sentry auth token (passed as --auth-token or the SENTRY_AUTH_TOKEN environment variable)
  • uv/uvx, or pip with Python, or Docker, to run the server
  • An MCP client such as Claude Desktop, VS Code or Zed

Setup

  1. Run with uvx (recommended)

    With uv, no specific installation is needed. uvx runs mcp-server-sentry directly. Add this to your claude_desktop_config.json.

    json
    "mcpServers": {
      "sentry": {
        "command": "uvx",
        "args": ["mcp-server-sentry", "--auth-token", "YOUR_SENTRY_TOKEN"]
      }
    }
  2. Install via pip

    Alternatively, install the package with pip.

    bash
    pip install mcp-server-sentry
  3. Run the pip-installed server

    After installation, run it as a script.

    bash
    python -m mcp_server_sentry
  4. Claude Desktop config with Docker

    Add this to your claude_desktop_config.json to run the server via Docker.

    json
    "mcpServers": {
      "sentry": {
        "command": "docker",
        "args": ["run", "-i", "--rm", "mcp/sentry", "--auth-token", "YOUR_SENTRY_TOKEN"]
      }
    }
  5. Zed configuration with uvx

    Add to your Zed settings.json.

    json
    "context_servers": [
      "mcp-server-sentry": {
        "command": {
          "path": "uvx",
          "args": ["mcp-server-sentry", "--auth-token", "YOUR_SENTRY_TOKEN"]
        }
      }
    ],

Examples

VS Code manual configuration (uvx)

json
json
{
  "mcp": {
    "inputs": [
      {
        "type": "promptString",
        "id": "auth_token",
        "description": "Sentry Auth Token",
        "password": true
      }
    ],
    "servers": {
      "sentry": {
        "command": "uvx",
        "args": ["mcp-server-sentry"],
        "env": {
          "SENTRY_AUTH_TOKEN": "${input:auth_token}"
        }
      }
    }
  }
}

What it does: Add to VS Code User Settings (JSON), or to .vscode/mcp.json in your workspace to share it. The mcp key is needed when using mcp.json. The token is prompted for as a password input.

Debug with the MCP inspector (uvx)

bash
bash
npx @modelcontextprotocol/inspector uvx mcp-server-sentry --auth-token YOUR_SENTRY_TOKEN

What it does: Uses the MCP inspector to debug a uvx installation of the server.

Debug from a local checkout

bash
bash
cd path/to/servers/src/sentry
npx @modelcontextprotocol/inspector uv run mcp-server-sentry --auth-token YOUR_SENTRY_TOKEN

What it does: For a package installed in a specific directory or under development.

Using the get_sentry_issue tool

Prompt
prompt
get_sentry_issue with issue_id_or_url set to a Sentry issue ID or URL

Expected output: The tool takes issue_id_or_url and returns the title, issue ID, status, level, first and last seen timestamps, event count and full stacktrace.

Pros & cons

Pros

  • Pro:Returns detailed issue data including the full stacktrace, status, level, timestamps and event count
  • Pro:Offers multiple run options: uvx, pip, and Docker
  • Pro:Documented configuration for Claude Desktop, VS Code and Zed
  • Pro:MIT licensed

Cons

  • Con:Exposes only one tool, which retrieves a single issue by ID or URL
  • Con:Lives in the servers-archived repository
  • Con:Zed uvx example in the README appears to use array syntax with key-value entries, which may need correction

Images