For Claude Code, Cursor & other agents
Set up Vitals in one prompt
No hunting for API keys. Your agent reads your codebase, works out which payment and analytics tools each product uses, walks you through creating read-only keys, connects everything, and shows you today's briefing.
1. Create an agent token
Start a free trial, then open Settings → Set up with your AI agent and click Create agent token.
2. Add Vitals to your agent
Claude Code:
claude mcp add --transport http vitals https://vitals-gold-six.vercel.app/api/mcp --header "Authorization: Bearer vt_your_token"
Cursor, Windsurf and other MCP clients (remote HTTP server):
{ "mcpServers": { "vitals": { "url": "https://vitals-gold-six.vercel.app/api/mcp", "headers": { "Authorization": "Bearer vt_your_token" } } } }3. Ask it to set things up
Set up Vitals for my products: find the payment and analytics providers each one uses, walk me through creating read-only keys, connect them, then show me today's briefing.
What your agent can do
- list_integrations
- Every source Vitals supports, the read-only key it needs, and exactly how to create it
- create_product / list_products
- Add the apps and sites you run, see health scores and connection status
- connect_source
- Send read-only keys. Vitals checks them with the provider, stores them encrypted, and loads 60 days of history
- list_resources / choose_resource
- Pick which PostHog project, Umami site or Vercel project to track
- get_briefing / get_metrics / list_alerts
- Ask how your products are doing without leaving the terminal
Safe by design
- Vitals only accepts read-only keys. A full Stripe secret key is rejected; your agent is given a link that creates a restricted, read-only one.
- Your agent is told to ask you before sending any key, and not to print secrets back.
- Keys are checked with the provider, then encrypted with AES-256-GCM before they're stored.
- Agent tokens can set up sources and read numbers, not change billing. Revoke one anytime in Settings.