Overview

Triw is an AI gateway. You call one endpoint with one key, and Triw routes the request to the provider that serves the model you asked for. Billing, quotas, keys and usage all live in a single console.

Base URL

https://router.triw.io

What you get

OpenAI-compatible API

POST /v1/chat/completions, /v1/embeddings, /v1/responses, /v1/images/generations, audio endpoints.

Anthropic-compatible API

POST /v1/messages and /v1/messages/count_tokens. Point Claude Code or the Anthropic SDK at Triw.

Gemini-compatible API

POST /v1beta/models/{model}:generateContent plus an OpenAI-style route at /v1beta/openai/models.

One bill

Every model is metered in the same quota unit, whichever provider serves it.

How a request flows

You send a request to https://router.triw.io with an API key (Authorization: Bearer sk-...).
Triw resolves the model name to an available upstream channel for your key’s group.
The upstream response is relayed back in the same wire format you sent.
Usage is metered against your balance.

You do not change your prompt or response parsing per provider — only the model value.

First request

curl https://router.triw.io/v1/chat/completions \
  -H "Authorization: Bearer $TRIW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

The complete route list — chat, messages, images, video, audio, embeddings, Midjourney tasks — is in Endpoints. Video jobs are task-based; see Video generation. New here? Read the quickstart. It takes about five minutes, from account to first token.

Getting started

Account

API

SDKs

Tools