Unified API
One OpenAI-compatible endpoint across all verified provider listings.
Buy or sell inference for Kimi, DeepSeek, Qwen, Llama and other open-weight models through one OpenAI-compatible API with pay-per-token pricing.

One OpenAI-compatible endpoint across all verified provider listings.
Route by lowest price, fastest latency, highest uptime or privacy policy.
Independent operators can register hosted open-weight models and sell inference.
Buyer requests run only while the account has a positive JohnRouter balance.
Compare price and provider coverage.
Loading models…
Add prepaid credits with Stripe and create an API key.
JohnRouter routes traffic to an eligible provider and meters token usage.
Providers register a model for $10, pass verification and set token prices.
Cleared seller earnings are paid once a week in arrears.