Hooklistener

Name: Hooklistener
Author: hooklistener

@hooklistenerMCP server · External

Webhook, email, WebSocket and tunnel testing for AI agents: capture, wait, verify, replay.

Software testingv1.0.1

Freeno charge per call

One address for any MCP client. You sign in at app.hooklistener.com once, after adding it.

What it does

Reads
Writes
High impact

At most: moves money, runs code or commands, deletes data, or sends messages for you.

What it costs

Free

Odel charges nothing per call. The publisher may charge for its own service.

What it asks for at sign-in

Access to	Read	Change
Full access	Not stated: full_access
Read only	Not stated: read_only

From the sign-in at app.hooklistener.com, as the endpoint publishes it. You approve it there.

Tools

None listed · checked 11 Oct 2026

Odel could not list its tools without signing in.

Try it in the Inspector

README4 min readView on GitHub

Hooklistener MCP Server

Hosted MCP server that lets AI coding agents test webhooks end to end: create a public webhook URL, wait for the webhook to arrive, verify its signature, and replay it to localhost. The same server also covers email inboxes, WebSocket/Socket.IO/MQTT/SSE endpoints, localhost tunnels and uptime monitors.

Endpoint: https://app.hooklistener.com/api/mcp
Transport: Streamable HTTP. Nothing to install or run locally.
Auth: OAuth 2.1 with PKCE and dynamic client registration. Clients request full_access or read_only. A Hooklistener API key (hklst_…) also works as a Bearer token for clients without OAuth.
Tools: 67, in 7 toolsets. A new workspace lists 26 of them (every webhook tool plus the tool that starts each other product); a toolset appears once you use that product. Send x-hooklistener-toolsets: all to list everything.
Plans: works on every plan, including Free.

This repository holds the setup instructions and the MCP Registry entry (server.json). The server itself is operated by Hooklistener.

Connect your client

Claude Code

bash

claude mcp add --transport http hooklistener https://app.hooklistener.com/api/mcp

Then run /mcp in Claude Code, select hooklistener and sign in with your browser. Full guide

Codex

bash

codex mcp add hooklistener \
  --url https://app.hooklistener.com/api/mcp \
  --oauth-resource https://app.hooklistener.com/api/mcp
codex mcp login hooklistener --scopes full_access   # or read_only

Full guide

Cursor

.cursor/mcp.json:

json

{
  "mcpServers": {
    "hooklistener": { "url": "https://app.hooklistener.com/api/mcp" }
  }
}

Full guide

VS Code (GitHub Copilot)

.vscode/mcp.json:

json

{
  "servers": {
    "hooklistener": { "type": "http", "url": "https://app.hooklistener.com/api/mcp" }
  }
}

Full guide

Claude.ai and Claude Desktop

Customize → Connectors → + → Add custom connector → paste https://app.hooklistener.com/api/mcp → Add → Connect. Full guide

Cline

Cline documents bearer-token headers for remote servers, so use a Hooklistener API key (paid plans, created at Organization Settings → API Keys). In cline_mcp_settings.json:

json

{
  "mcpServers": {
    "hooklistener": {
      "type": "streamableHttp",
      "url": "https://app.hooklistener.com/api/mcp",
      "headers": { "Authorization": "Bearer hklst_your_api_key" }
    }
  }
}

Agents installing the server for you can follow llms-install.md.

Other clients

Step-by-step guides: ChatGPT · Gemini CLI · Windsurf · Zed · OpenCode · Grok (xAI API)

Any MCP client that supports Streamable HTTP can connect with the endpoint above.

What your agent can do

Category	Examples
Debug endpoints	`create_endpoint`, `get_endpoint`, `list_endpoint_anomalies`, `set_endpoint_alerts`
Captured requests	`wait_for_request`, `list_requests` (list or search), `get_request`, `diagnose_request`, `investigate_request_retries`
Verify, compare and replay	`verify_request_signature` (Stripe, GitHub, Slack), `validate_request` (JSON Schema), `diff_requests`, `replay_request` (replay or forward, edit the body and re-sign)
Mock responses	`create_response_rule`, `test_response_rules`
Request threads	`set_thread_rule`, `list_request_threads`
Replay cases and suites	`save_request_case`, `run_endpoint_cases`, `wait_for_case_run`
Real-time endpoints	`create_realtime_endpoint` (WebSocket, Socket.IO, MQTT, SSE), `wait_for_realtime_message`, `send_realtime_message`
Email inboxes	`create_inbox`, `wait_for_email`, `get_email`
Localhost tunnels	`plan_tunnel_action`, `replay_tunnel_capture`
Uptime monitors	`create_monitor`, `get_monitor_status`
Secrets and tasks	`create_secret`, `cancel_task`

Every tool has a title and read-only/destructive annotations. Full reference: docs.hooklistener.com/mcp/available-tools

Example prompts

"Create a Hooklistener endpoint for Stripe, trigger a test checkout, and wait for checkout.session.completed."
"Verify that webhook's signature with my Stripe secret, then replay it to http://localhost:3000/webhooks, re-signed."
"Create an inbox, sign up with it on staging, and read me the verification code from the email."
"Save the last three webhooks as a replay suite and run it against localhost after my fix."

Safety

Read-only access: request the read_only scope and the agent can search, wait for, diff, validate and diagnose traffic, but can't create, edit, delete, replay, forward or send anything.
Deletes are two-step: preview with dry_run, then confirm.
Replays and forwards need an idempotency key, so a retried call never sends a webhook twice.
Credentials are masked in tool results: common secrets in headers, cookies, query strings and JSON bodies.
Data is scoped to your organization and hosted in Europe.