Hookline MCP

Hookline MCP connects your agent to Hookline, a tool for inspecting, mocking and validating webhooks and API payloads. Your agent can run the whole loop without leaving the editor: create an endpoint, wait for a request, validate the payload, read the findings and adjust the mock response. What you can do with it - Create test webhook endpoints, optionally set to expire, and get the URL to send requests to - Wait for the next request after a send, and read requests with their headers, query strings and bodies - Validate payloads against your Hookline validators, which cover built-in specs such as OpenRTB, CloudEvents and Standard Webhooks, or your own JSON Schema - Compare two payloads, search rules and read validation history - Write custom rules by field path, create validators and set an endpoint's mock response - Make a read-only, expiring share link to a single received request How it connects It runs on your machine with npx -y @hookline/mcp and needs Node 20 or later. It takes a Hookline agent key (hlk_...) in HOOKLINE_API_KEY, which you create at app.hookline.org/settings/mcp. Hookline has a free plan for one workspace. Read-only by default Hookline recommends the read-only key preset as the safe default. With a read-only key the write tools aren't registered at all, so your agent never sees them. Creating endpoints, validators, rules, share links and mock responses needs a key with write access. Good to know Webhook bodies come from outside senders, so the server marks them as untrusted data. Hookline says plainly that this lowers the risk of prompt injection but doesn't remove it. Treat mock response bodies as public: anyone who knows an endpoint's URL can read them.