◆ ONEPROMPT

One API.
Every model, triaged at the edge.

An OpenAI- and Anthropic-compatible endpoint. A fast edge model reads every request, answers the simple ones instantly, and escalates only the hard ones — so you don't pay a frontier model to say hello. MCP-tool-ready.

Dual API

Point an OpenAI or Anthropic client at it — /v1/chat/completions and /v1/messages both work.

Edge triage

A Workers-AI router picks edge vs deep per call. Simple stays at the edge; hard escalates.

MCP-ready

Tools run at the edge when they can fully answer — the upstream model is only troubled when it's actually needed.

curl https://oneprompt.site/v1/chat/completions \
  -H "authorization: Bearer $ONEPROMPT_KEY" \
  -H "content-type: application/json" \
  -d '{"model":"oneprompt-auto","messages":[{"role":"user","content":"Hello"}]}'
Models: oneprompt-auto · oneprompt-edge · oneprompt-deep — see /v1/models · /health. Beta — a cinematic landing + live MCP tools are in flight.