Skip to Content

@rustrak/mcp

@rustrak/mcp is a Model Context Protocol  server for Rustrak. Connect it to Claude Desktop, Cursor, Continue.dev, or any MCP-compatible AI assistant and get 18 tools to manage your error tracking without leaving your AI tool.

npx @rustrak/mcp # run directly with npx (no install needed) # or install globally npm install -g @rustrak/mcp

Requirements: Node.js ≥ 18

How It Works

The server runs as a local stdio process — no open network port, no daemon. Your AI client starts it automatically when needed and communicates via JSON-RPC over stdin/stdout. The Rustrak API token is loaded from environment variables and never exposed as a tool argument.

AI Client (Claude / Cursor / Continue) │ stdio (JSON-RPC) ┌─────────────────────┐ │ @rustrak/mcp │ │ ├── projects │ │ ├── issues │ │ ├── events │ │ ├── tokens │ │ └── alerts │ └──────────┬──────────┘ │ HTTP via @rustrak/client Rustrak Server

Setup

1. Create an API token

In the Rustrak web UI: Settings → Tokens → Create token. Save the full token value — it is shown only once.

2. Configure your AI client

Claude Desktop~/Library/Application Support/Claude/claude_desktop_config.json:

{ "mcpServers": { "rustrak": { "command": "npx", "args": ["-y", "@rustrak/mcp"], "env": { "RUSTRAK_API_URL": "https://your-rustrak-instance.example.com", "RUSTRAK_API_TOKEN": "your-40-char-hex-token" } } } }

Cursor.cursor/mcp.json (project-level) or ~/.cursor/mcp.json (global):

{ "mcpServers": { "rustrak": { "command": "npx", "args": ["-y", "@rustrak/mcp"], "env": { "RUSTRAK_API_URL": "https://your-rustrak-instance.example.com", "RUSTRAK_API_TOKEN": "your-40-char-hex-token" } } } }

Continue.dev~/.continue/config.json:

{ "experimental": { "modelContextProtocolServers": [ { "transport": { "type": "stdio", "command": "npx", "args": ["-y", "@rustrak/mcp"], "env": { "RUSTRAK_API_URL": "https://your-rustrak-instance.example.com", "RUSTRAK_API_TOKEN": "your-40-char-hex-token" } } } ] } }

Environment Variables

VariableRequiredDescription
RUSTRAK_API_URLBase URL of your Rustrak server
RUSTRAK_API_TOKEN40-char hex API token

The server exits immediately with a clear error message if either variable is missing.

Available Tools (18)

Projects

ToolDescription
list_projectsList all projects
get_projectGet project details including DSN
create_projectCreate a new project

Issues

ToolDescription
list_issuesList issues with filters (open / resolved / muted / all)
get_issueGet a single issue with full details
resolve_issueMark an issue as resolved
unresolve_issueRe-open a resolved issue
mute_issueMute an issue (silences future alerts)
delete_issue⚠️ Permanently delete an issue and all its events

Events

ToolDescription
list_eventsList raw events for an issue (cursor pagination)
get_eventGet a single event with full Sentry envelope data

Tokens

ToolDescription
list_tokensList API tokens (masked)
create_tokenCreate a new API token (full value shown once)
revoke_token⚠️ Permanently revoke an API token

Alerts

ToolDescription
list_alert_channelsList notification channels (Slack, email, webhook)
test_alert_channelSend a test notification to a channel
list_alert_rulesList alert rules for a project

Tools marked ⚠️ are annotated with destructiveHint — supported clients will prompt for confirmation before executing.

Example Prompts

Once connected, you can ask things like:

  • “List all unresolved issues in project 1”
  • “Show me the full stack trace for issue abc-123”
  • “Resolve all TypeError issues from the last deployment”
  • “Create a token called ‘CI pipeline’ and give me the value”
  • “How many events does issue xyz have?”
  • “Mute all database timeout issues”

Security Notes

  • The API token is loaded exclusively from RUSTRAK_API_TOKEN — it is never accepted as a tool argument
  • All tool errors are returned as isError: true content rather than exceptions, so your AI can see and handle them
  • The server uses stdio transport — it never opens a network port
Last updated on