Who it's for
If HTTP traffic is your product, EchoRelay is the layer in the middle — the target your callers hit and the source that calls your services. Here is who reaches for it, and where each one starts.
For AI builders
You are wiring an LLM or inference service to your users — and losing days to streaming, auth, and rate-limits instead of the model.
Point your streaming target at EchoRelay and we forward its SSE or chunked response to your caller, chunk by chunk. Set the whole endpoint up by telling your agent — it speaks our MCP surface and ships the config for you.
See streamingFor founders & small teams
Every hour on webhook logic, API keys, rate-limits, and retry queues is an hour not on your product — and you have no platform team to hand it to.
EchoRelay is the middle layer: your endpoint up front, with keys, limits, and validation handled for you. Free to start, no card, and configurable from a single Claude session.
See origin shieldFor event-heavy teams
You fan a webhook or event out to a dozen downstream services, and a dropped delivery costs you a transaction — so you maintain retry logic, dead-letter queues, and per-consumer limits yourself.
One request fans out to every target in parallel — delivered at-least-once, retried automatically, with a dead-letter queue you can inspect and replay. Nothing is silently dropped.
See fan-outWhichever you are: your traffic runs on infrastructure we operate in the EU, every knob is reachable over an API and an MCP server, and the free tier applies from request one — no card. Not sure which shape fits? See all four relay shapes, or browse by use case.
No credit card. Configure it yourself or hand it to your agent.