MCP Directory

Sentry MCP Server (remote)

Official

Official hosted Sentry MCP for investigating issues, errors, traces and performance data.

Verified
HTTP (remote)
OAuth
TypeScript

Add to your client

Copy the config for your MCP client and paste it into its config file.

Install / run
Connect client to https://mcp.sentry.dev/mcp and complete OAuth

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "sentry-mcp-server-remote": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.sentry.dev/mcp"
      ]
    }
  }
}

Claude Desktop connects to remote servers through the `mcp-remote` proxy (installed on first run via npx). Restart Claude Desktop after saving.

Step-by-step guides: Add to Claude Desktop · Add to Cursor · Add to Windsurf

Before you start

  • A Sentry account (SaaS or self-hosted) with access to the org/projects you want to query
  • For the hosted server: an MCP client that supports remote MCP over HTTP with OAuth (the OAuth flow runs in your browser)
  • For the local stdio variant: Node.js plus a Sentry User Auth Token with scopes such as org:read, project:read/write, team:read/write, and event:write
  • Seer must be enabled on your Sentry org to use the AI root-cause analysis tool

About Sentry MCP Server (remote)

The official Sentry MCP server, maintained by Sentry, lets AI assistants investigate your Sentry data: issues, errors, traces, performance/spans, releases, projects, teams, and DSNs. It is primarily offered as a hosted remote server at https://mcp.sentry.dev using OAuth, so there are no API keys to paste into your client config.

Beyond simple lookups, it includes AI-powered tools that translate natural-language questions into Sentry's query syntax (search_events, search_issues) and an analyze_issue_with_seer tool that invokes Sentry's Seer agent for root-cause analysis. It also exposes documentation search so the assistant can ground answers in Sentry's docs.

The same project ships a stdio variant (@sentry/mcp-server) that you can run locally with a user auth token if you prefer a local process or are connecting a client that does not yet support remote OAuth MCP servers.

Tools & capabilities (12)

find_organizations

List the Sentry organizations the user can access

find_projects

List projects within an organization

get_issue_details

Fetch full details for a specific issue by ID or URL

search_issues

Natural-language search that resolves to Sentry issue queries

search_events

Natural-language search over events/errors/spans

get_trace_details

Retrieve a distributed trace by ID

analyze_issue_with_seer

Run Sentry's Seer agent for AI root-cause analysis on an issue

update_issue

Change an issue's status or assignment

find_releases

List releases for a project or organization

create_project

Create a new Sentry project (also create_team, create_dsn)

search_docs

Search Sentry's documentation; get_doc fetches a page

whoami

Return the authenticated user's identity

When to use it

  • Use it when an exception shows up in Sentry and you want the assistant to pull the stack trace and surrounding context into your editor
  • Use it when you want plain-English queries over errors and events instead of writing Sentry search syntax by hand
  • Use it when you need AI root-cause analysis on a tricky issue via Sentry's Seer
  • Use it when correlating a slow request with its distributed trace and spans
  • Use it when triaging: listing recent issues for a project and updating their status or assignee from chat

Quick setup

  1. 1Add the remote server URL `https://mcp.sentry.dev/mcp` to your MCP client (or install via the client's connector UI)
  2. 2Complete the OAuth flow in your browser to grant access to your Sentry org
  3. 3Approve the requested scopes for the projects you want the assistant to read
  4. 4Restart or reconnect your MCP client so the Sentry tools load
  5. 5Verify by asking the assistant to list your organizations or recent issues
  6. 6Alternatively, run `npx @sentry/mcp-server@latest --access-token=...` locally for stdio mode

Security notes

The OAuth grant exposes your Sentry org's error and trace data, which can contain PII captured in events; review scopes before approving. For the stdio fallback, treat SENTRY_ACCESS_TOKEN as a secret.

Sentry MCP Server (remote) FAQ

Is the Sentry MCP server official?

Yes. It is built and maintained by Sentry (getsentry/sentry-mcp) and is licensed Apache-2.0.

Do I need an API key for the hosted server?

No. The hosted server at mcp.sentry.dev uses OAuth, so you authenticate in the browser rather than pasting a token. A user auth token is only needed for the local stdio variant.

Remote vs local: which should I use?

Use the hosted remote server for the simplest setup and OAuth auth. Use the local @sentry/mcp-server stdio variant if your client lacks remote-OAuth support or you want a self-contained process.

What is the analyze_issue_with_seer tool?

It triggers Sentry's Seer AI agent to perform automated root-cause analysis on an issue. Seer must be enabled for your organization.

Does it work with self-hosted Sentry?

The hosted server targets Sentry SaaS; for self-hosted instances you typically run the stdio server locally configured against your instance and a user auth token.

Alternatives to Sentry MCP Server (remote)

Compare all alternatives →

Microsoft's official browser-automation MCP using Playwright's accessibility tree (no vision model).

Featured
Verified
stdio (local)
No auth
TypeScript
12 tools
Updated 9 days agoRepo

Up-to-date, version-specific library documentation injected into your coding agent.

Verified
stdio (local)
API key
TypeScript
2 tools
Updated 13 days agoRepo

LSP-powered coding agent toolkit: semantic symbol search, references and structural edits.

Verified
stdio (local)
No auth
Python
11 tools
Updated 11 days agoRepo