Allows AI coding agents to debug Stripe webhooks by syncing and listing Stripe events, retrieving full event payloads, waiting for specific event types, and replaying events with valid signatures into a local handler while returning the handler's response status and body.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@stripe-webhook-mcp invoice.paid is 500ing — replay the latest one and show the error"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Let your AI coding agent debug Stripe webhooks itself. Claude Code or Cursor can pull Stripe events, read their payloads, and replay them, correctly signed, into your local handler, then read back the status and stack trace. No dashboard tab, no copy-pasting JSON.
you: "invoice.paid is 500ing, fix it"
agent: list_events(type: "invoice.paid") → replay_event(evt_…) → 500, TypeError at webhook.ts:59
…edits handler… → replay_event(evt_…) → 200 ✓You need a test-mode Stripe key, plus the signing secret your handler verifies with.
Claude Code
claude mcp add stripe-webhooks \
-e STRIPE_SECRET_KEY=sk_test_... \
-e STRIPE_WEBHOOK_SECRET=whsec_... \
-e WEBHOOK_TARGET_URL=http://localhost:3000/api/webhooks/stripe \
-- npx -y stripe-webhook-mcp mcpCursor (.cursor/mcp.json)
{
"mcpServers": {
"stripe-webhooks": {
"command": "npx",
"args": ["-y", "stripe-webhook-mcp", "mcp"],
"env": {
"STRIPE_SECRET_KEY": "sk_test_...",
"STRIPE_WEBHOOK_SECRET": "whsec_...",
"WEBHOOK_TARGET_URL": "http://localhost:3000/api/webhooks/stripe"
}
}
}
}A restricted key (rk_test_...) with read access to Events is enough. Live keys are
refused unless you set STRIPE_WEBHOOK_MCP_ALLOW_LIVE=1.
Related MCP server: webhook-ai-mcp
Tool | What it does |
| Recent events as one-line summaries. Filter by type ( |
| The full payload of one event, plus its replay history |
| Block until an event type arrives. Use it after triggering a test checkout |
| Re-deliver an event, signed, to your handler, and return its status and body |
| Past replays and what the handler returned |
| Pull new events from Stripe. The other tools call it as needed |
Events come from Stripe's Events API using your key, so there's no tunnel and no
stripe listen to run. They're stored in a local SQLite file
(~/.stripe-webhook-mcp/events.db). Replays are signed with your whsec_ the same way
Stripe signs them, so your handler's constructEvent check passes unmodified.
The same operations are available without an agent:
npx stripe-webhook-mcp sync
npx stripe-webhook-mcp list --type invoice.paid
npx stripe-webhook-mcp replay evt_123 --to http://localhost:3000/api/webhooks/stripeexamples/webhook-fixture.ts is a small handler with a deliberate, realistic bug in
invoice.paid:
STRIPE_WEBHOOK_SECRET=whsec_test npm run fixture # http://localhost:3000/webhookMIT
This server cannot be deployed
Debug webhooks from your AI agent: inspect and replay captured webhooks on localhost.
Webhook, email, WebSocket and tunnel testing for AI agents: capture, wait, verify, replay.
Receive, inspect, replay and deliver webhooks — with signature verification and agent triggers.
Anonymous webhook capture, inspection, waiting, and response configuration for AI agents.
