@rustrak/mcp

Manage a self-hosted Rustrak error tracker: issues, events, traces, alerts, tokens and team

Official · Rustrak TypeScript Local · stdio DevOps and monitoring

0 popularity score ★ 846 ↓ 61/wk

Connect your client

Local (npm)

Add in one click

Opens the app, which asks before it adds the server. Then replace each <NAME> in its settings with your value.

Or add it by command or config

It needs RUSTRAK_API_URL, RUSTRAK_API_TOKEN. Replace each <NAME> with your value.

Claude CodeTerminal
claude mcp add rustrak -e RUSTRAK_API_URL=<RUSTRAK_API_URL> -e RUSTRAK_API_TOKEN=<RUSTRAK_API_TOKEN> -- npx -y @rustrak/mcp
Claude Desktopclaude_desktop_config.json
{
  "mcpServers": {
    "rustrak": {
      "command": "npx",
      "args": [
        "-y",
        "@rustrak/mcp"
      ],
      "env": {
        "RUSTRAK_API_URL": "<RUSTRAK_API_URL>",
        "RUSTRAK_API_TOKEN": "<RUSTRAK_API_TOKEN>"
      }
    }
  }
}
Cursor~/.cursor/mcp.json
{
  "mcpServers": {
    "rustrak": {
      "command": "npx",
      "args": [
        "-y",
        "@rustrak/mcp"
      ],
      "env": {
        "RUSTRAK_API_URL": "<RUSTRAK_API_URL>",
        "RUSTRAK_API_TOKEN": "<RUSTRAK_API_TOKEN>"
      }
    }
  }
}
VS Code.vscode/mcp.json
{
  "servers": {
    "rustrak": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@rustrak/mcp"
      ],
      "env": {
        "RUSTRAK_API_URL": "<RUSTRAK_API_URL>",
        "RUSTRAK_API_TOKEN": "<RUSTRAK_API_TOKEN>"
      }
    }
  }
}
Codex~/.codex/config.toml
[mcp_servers.rustrak]
command = "npx"
args = ["-y", "@rustrak/mcp"]

[mcp_servers.rustrak.env]
RUSTRAK_API_URL = "<RUSTRAK_API_URL>"
RUSTRAK_API_TOKEN = "<RUSTRAK_API_TOKEN>"
Gemini CLITerminal
gemini mcp add --scope user -e RUSTRAK_API_URL=<RUSTRAK_API_URL> -e RUSTRAK_API_TOKEN=<RUSTRAK_API_TOKEN> rustrak npx -- -y @rustrak/mcp

Settings 2

RUSTRAK_API_URL Required: Base URL of your Rustrak server; the server exits without it.
RUSTRAK_API_TOKEN Required: 40-character hex API token from Settings → Tokens; the server exits without it.

What it does

@rustrak/mcp is the MCP server published by the Rustrak project for its own self-hosted, Sentry-compatible error tracker. It runs on your machine as a stdio process and talks to your Rustrak instance over its REST API through the @rustrak/client package, using an API token you create in the Rustrak web UI.

With it an assistant can browse projects, list and triage issues (resolve, re-open, mute, assign, change status or priority, comment, delete singly or in bulk), read raw events with their full Sentry payload, and look at performance data: transactions with p50/p95/p99 latency, spans, logs, release health and project-wide error volume. It also exposes the AI agent monitoring views (agent runs, LLM calls and tokens by model, tool calls, agent traces), and administration: API tokens, alert channels and rules, team members, invitations, per-project roles and storage cleanup.

The tool count differs between its own pages: the npm README says 18 tools, the docs page lists 27, and the source at tag v0.16.4 (packages/mcp/src/server.ts and the files under tools/) registers 68. The extra ones cover transactions, spans, logs, release health, stats, storage, AI agent monitoring, a server version check and more issue actions.

What it needs

RuntimeNode.js 18+

Node.js 18 or later (the launch command uses npx). A running Rustrak server you can reach, self-hosted by you, and an API token created in its web UI under Settings → Tokens. Team, invitation and storage tools need the token to belong to an admin user (or be a legacy full-access token); project-member tools need global-admin or project-admin access.

Limits

It only works against a Rustrak server; it does not talk to Sentry or other trackers directly. The server exits at start if RUSTRAK_API_URL or RUSTRAK_API_TOKEN is missing (config.ts). list_issues returns open issues by default and at most 100 per page with per_page. execute_storage_cleanup and gc_storage_source_maps refuse unless called with confirm=true and allow one cleanup at a time; bulk_delete_issues has no such confirmation and with filter=all deletes every issue in a project. The docs say a token value is shown only once, but the get_token tool returns a token's full 40-character value by ID.

Who it suits

Teams that already run Rustrak and want to triage errors, check performance or release health, or administer the instance from Claude, Cursor or Continue without opening the web UI. Because many tools change or permanently delete data and a token can read other tokens' values, it suits people who can give it a token scoped to what they want the assistant to do; an admin token unlocks team and storage tools. If you use hosted Sentry or another tracker instead of Rustrak, this server is of no use to you.

Tools 60 of 68

list_projects
List the Rustrak projects the token can see
get_project
Show one project's details, including its DSN
create_project
Create a new project
list_issues
List a project's issues, filtered by state (open by default)
get_issue
Fetch one issue with its details
resolve_issue
Mark an issue resolved
unresolve_issue
Re-open a resolved issue
mute_issue
Silence an issue so it stops raising alerts
delete_issue
Permanently remove an issue and its events
update_issue_status
Set an issue's status and optionally its priority
assign_issue
Assign an issue to a user or clear the assignee
bulk_update_issues
Change status or priority of many issues at once
bulk_delete_issues
Permanently delete issues by ids or by filter, up to every issue in a project
get_issue_stats
Event-count timeseries for one issue over 24h or 30d
get_issue_activity
Show an issue's activity history
comment_on_issue
Add a note to an issue
submit_user_report
Attach a user feedback report to an issue
list_events
List an issue's raw events with cursor paging
get_event
Fetch one event with its full Sentry envelope
list_tokens
List API tokens with values masked
get_token
Return one token's full 40-character value
create_token
Create a new API token
revoke_token
Permanently revoke an API token
list_alert_channels
List notification channels: Slack, email, webhook, custom webhook
create_alert_channel
Add a notification channel, with a body template for custom webhooks
update_alert_channel
Rename, toggle or change credentials of a channel
preview_alert_template
Render a custom webhook body against a sample alert
test_alert_channel
Send a test notification through a channel
list_alert_rules
List a project's alert rules
list_team_members
List instance users with role and status
update_member_role
Switch a user between admin and member
remove_team_member
Remove a user from the instance
create_invitation
Invite someone by email and get a one-time invite token
list_invitations
List pending, accepted and revoked invitations
revoke_invitation
Cancel a pending invitation
list_project_members
List a project's members and their roles
set_project_member
Add a user to a project or change their role
remove_project_member
Take away a user's access to a project
get_server_version
Report the Rustrak server version
get_release_health
Crash-free session and user rates per release
list_transactions
List performance transactions, newest first, with filters
get_transaction
Fetch one transaction with spans, trace context and web vitals
get_transaction_stats
Throughput, p50/p95/p99 latency and failure rate per transaction
get_transaction_spans
Spans of one transaction in waterfall order with timings
list_logs
List a project's logs, filtered by level or trace
list_spans
List spans, optionally only AI agent span types
get_error_volume
Error events over time, bucketed by severity
get_project_stats
Headline counts for a window compared with the window before
get_agent_runs
Time-bucketed count of AI agent runs
get_agent_duration
Average and p95 duration of agent runs and LLM calls
get_agent_models_by_calls
Top LLM models by number of calls
get_agent_models_by_tokens
Top LLM models by tokens used
get_agent_tools
Most-called tools in agent traces
list_agent_traces
List AI agent traces with duration, tokens and tool calls
get_storage_summary
Instance-wide database size and row counts (admin)
get_storage_by_project
Storage use broken down per project (admin)
preview_storage_cleanup
Dry run of a retention cleanup showing what would go
execute_storage_cleanup
Delete data older than N days; needs confirm=true (admin)
get_storage_cleanup_status
Progress or result of the latest cleanup
gc_storage_source_maps
Delete orphaned source-map files; needs confirm=true (admin)

Similar servers