Developers
Read it from your terminal, your CI and your agent.
A read-only API, an MCP server, a CLI and a CI upload. A token only reads, and never sees more than the person who made it can see now.
An MCP server for your agent
Claude Code, Claude Desktop, Cursor and other agents answer questions about your projects: what expires this month, what broke since the last deploy, which project lost money.
- Streamable HTTP, stateless, ten read-only tools.
- Without amounts, answers leave the money out and say so.
claude mcp add --transport http shipvitals https://shipvitals.com/api/mcp \
--header "Authorization: Bearer sv_api_…"
A read-only API
The portfolio's figures, one project's, costs by provider or service, revenue and MRR movements, renewals due, uptime and incidents, as JSON. The reference is drawn from the OpenAPI 3.1 document the API serves.
- 60 requests a minute per token; 429 with Retry-After over it.
- Each request runs in a read-only database transaction.
curl https://shipvitals.com/api/v1/public/projects \
-H "Authorization: Bearer sv_api_…"
The vitals CLI
The same engine as the desktop app: it runs a project's own checks on any machine with Node 24 and says whether the project can ship. It works without an account for local results.
- vitals doctor says what this machine can run.
- vitals agent keeps a headless box running checks.
vitals doctor
vitals link .
vitals checks approve
vitals run
Also in ShipVitals
Tokens scoped like a person
A workspace token covers the projects its maker chose, narrows as soon as their access narrows and stops when they leave. ShipVitals stores only its SHA-256 hash and shows it once.
CI results and your own metrics
Your CI uploads its test results, and your servers push the metrics no service reports, such as active users. Both are set up per project.
Every scope, on one page
What ShipVitals asks each of its connectors for, and which keys aren't read-only, from the same definitions the connect dialogs show.