Monitor & debug · official server

Sentry for your AI agent

Give Claude Code, Cursor, Codex or your own agent Sentry through one MCP URL and one key. Your team signs in to Sentry once and the agent never holds a token. Reads run at once; anything that changes Sentry waits for your approval.

210
Sentry actions your agent gets as tools
1 URL
one endpoint for every assistant
60/min
calls per key, hard caps on top
0
Sentry credentials inside your agent

Actions

What your agent can do in Sentry

A sample of the 210 actions. Your agent sees the full list with tools/list.

Reads · run at once

  • Access project information
  • Fetch alerts
  • Fetch a monitor
  • Fetch an alert
  • Fetch an organization's monitors
  • Fetch issue event by id
  • Fetch organization alert rules
  • Fetch organization release threshold statuses

Writes · wait for your approval

  • Add a symbol source to a project
  • Add organization member via email
  • Add or remove user email by id
  • Add team member in organization
  • Add team to project
  • Bulk mutate an organization's issues
  • Bulk remove an organization's issues
  • Create an alert for an organization

Official server

Also: Sentry's own MCP server

Sentry publishes its own MCP server at https://mcp.sentry.dev/mcp. The workspace owner connects it on the Connections page with Sentry's own sign-in; Hub keeps that sign-in encrypted and passes Sentry's tools through as sentry__<tool>. A tool Sentry marks read-only runs at once; every other one waits for your approval and runs once.

  1. Create your Vellarin account (free, no card).
  2. Open Connections → Official servers and choose Sentry. Its own sign-in page opens; approve with your account.
  3. Add Vellarin Hub to your assistant once (no code for Claude and ChatGPT; one line for Claude Code, Cursor or Codex).

Quickstart

Add Hub to your agent

  1. Create your account and connect Sentry with its normal sign-in page.
  2. Open Developers and create an agent key (you can limit it to Sentry).
  3. Add the server to your client:

Claude Code · Terminal

claude mcp add --transport http vellarin https://hub.vellarin.ai/api/hub/mcp --header "Authorization: Bearer <key>"

Cursor · ~/.cursor/mcp.json (or .cursor/mcp.json in a project)

{
  "mcpServers": {
    "vellarin": {
      "url": "https://hub.vellarin.ai/api/hub/mcp",
      "headers": {
        "Authorization": "Bearer ${env:VELLARIN_HUB_KEY}"
      }
    }
  }
}

Codex · ~/.codex/config.toml

[mcp_servers.vellarin]
url = "https://hub.vellarin.ai/api/hub/mcp"
bearer_token_env_var = "VELLARIN_HUB_KEY"

Questions

Sentry and AI agents

How do I connect Sentry to Claude Code, Cursor or Codex?

Create a free Vellarin account, connect Sentry with its normal sign-in, create an agent key on the Developers page and add the Vellarin Hub MCP server to your client with one line. Your agent then sees the Sentry actions as tools.

Can my agent change things in Sentry without asking me?

No. Read actions run at once. Any action that sends, creates, changes or deletes something in Sentry waits for the workspace owner's approval and then runs exactly once.

Does my agent get my Sentry password or token?

No. The sign-in lives on the server side. Your agent holds only a Vellarin Hub key or sign-in, which you can scope, cap per day and per month, and revoke at any moment.

What does it cost?

Hub Free includes 500 app calls a month and 1 agent key, with no card. Over the cap the answer is "upgrade or wait", never a charge. Your Sentry plan is between you and Sentry.

More behind the same address

All 104 →

Sentry is a trademark of its owner. Vellarin Hub is an independent product and is not affiliated with or endorsed by it.