Overview

One OpenAI-compatible API for 80+ open-weight models, called from Morse's attested enclave, which keeps no prompt or reply.

Morse is an OpenAI-compatible API for 85 open-weight chat models, from DeepSeek, Kimi, and GLM to Qwen, MiniMax, and gpt-oss. Morse calls every provider from an attested enclave that keeps no prompt or reply and never tells a provider who you are.

If your code already calls OpenAI, it calls Morse by changing the base URL, the key, and the model.

curl https://inference.trymorse.ai/v1/chat/completions \
  -H "Authorization: Bearer $MORSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshotai/kimi-k3.1-pro",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Start here

QuickstartMake a key, then send your first request with curl, Python, or TypeScript.AuthenticationSend your key as a Bearer token, keep it on a server, and limit what each key can do.ModelsModel ids, how to list them, and what each model can do on each endpoint.API referenceThe base URL, authentication, the request and reply as types, finish reasons, and cost.

What makes Morse different

Privacy tiersPick how private each request is, from a provider under zero retention, to secure hardware no one can read.Provider routingSet a privacy floor on a request or a key. Morse refuses a request it can't serve at that floor rather than sending it lower.Verification and attestationCheck that Morse's enclave runs the code Morse publishes before you send a prompt.StreamingStreaming, tool calling, and reasoning work as they do with OpenAI. Tools are limited on some encrypted endpoints.

Base URL

Every endpoint lives under one address:

Base URL

https://inference.trymorse.ai/v1

It speaks OpenAI's Chat Completions format. See Switch from OpenAI or OpenRouter for what carries over unchanged and what differs.