About
Hey Tiny Launch! π
I built HookWatch after a Stripe webhook silently failed for 3 days. Customers were paying but never getting access. No error. No alert. Just angry support tickets.
That's the thing about webhooks, cron jobs, WebSocket connections, and AI agent calls β when they break, they break silently. No stack trace, no error page. You find out when it's already too late.
HookWatch gives you four monitoring tools under one roof:
π Webhook Monitor β Track every incoming webhook in real-time. Inspect payloads, replay failed deliveries, get automatic retries with exponential backoff. Unique feature: if your server goes down, HookWatch buffers incoming webhooks and replays them when you're back online. Zero data loss.
β° Cron Monitor β Write schedules in plain English ("every day at 2am") instead of decoding 0 2 * * *. Full execution history with stdout/stderr, automatic retries, and instant alerts when something breaks.
π WebSocket Monitor β A transparent proxy that captures every message in both directions. Point your client to a HookWatch proxy URL β we forward everything while logging full traffic. Debug dropped connections and inspect payloads without changing a single line of code.
π€ MCP Proxy β Observability for AI agent tool calls. Monitor every MCP request/response, track latency (p50/p95/p99), and get alerts when your agents hit errors. If you're building with Claude or other LLMs, this gives you the visibility you've been missing.
What makes it different:
Four tools, one dashboard β competitors focus on just one
Local-first CLI that works offline β not just a web UI
Transparent proxying for webhooks and WebSockets β zero code changes
Human-readable cron schedules β stop Googling cron syntax
AI-native from day one β MCP observability built in, not bolted on
Built as a solo dev from Moldova π²π© β would love your feedback! What features or integrations would you want to see?
π hookwatch.dev