Conway

Overview

Multi-provider inference API. Run frontier models billed from Conway credits.

Conway Compute is a multi-provider inference proxy with an OpenAI-compatible interface. Send chat completions to OpenAI, Anthropic, Google, Moonshot, or Qwen models through a single endpoint, billed from your Conway credits.

POST https://inference.conway.tech/v1/chat/completions

How It Works

All requests go through POST /v1/chat/completions with an OpenAI-compatible payload. Conway routes to the right provider based on the model name and returns a normalized OpenAI-format response.

curl -X POST https://inference.conway.tech/v1/chat/completions \
  -H "Authorization: Bearer cnwy_k_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4.5",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

Supported Models

ProviderModels
OpenAIgpt-5.2, gpt-5.2-codex, gpt-5-mini, gpt-5-nano
Anthropicclaude-opus-4.6, claude-opus-4.5, claude-sonnet-4.5, claude-haiku-4.5
Google Geminigemini-2.5-pro, gemini-2.5-flash, gemini-3-pro, gemini-3-flash
Moonshotkimi-k2.5
Qwenqwen3-coder

Features

  • Single endpoint -- one API for all providers
  • OpenAI-compatible -- drop-in replacement for OpenAI SDK
  • Streaming -- SSE streaming for all providers
  • Credit-based billing -- pay from your Conway balance, no per-provider API keys needed
  • x402 payments -- permissionless pay-per-request with USDC (no account needed)

See the full API reference for request parameters, response format, and billing details.