Privacy-first long-form inference: 8,192-token output cap on a local open-weight model on dedicated hardware - prompts never reach OpenAI/Anthropic or any third-party cloud. OpenAI-compatible chat completions (messages array in, chat.completion JSON out) for summarization, drafting, extraction, translation, code, multi-paragraph answers. Pay per call in USDC on Base (x402), no API key or account. Shorter work? Use the cheaper quick tier (4,096 tokens).
| Network | Scheme | Amount | Pay To |
|---|---|---|---|
| Base | exact | $0.05 USDC | 0x2397...1ee8 |