An unmetered LLM API

This service is a flat rate LLM API built for coding agents. It serves the Qwen3.8 Flash model through an OpenAI compatible chat completions endpoint, so any client that accepts a custom base URL, API key and model ID can use it. Instead of billing per token, it charges a fixed monthly subscription, removing the variable usage bill that long running agent sessions tend to create. It is aimed at developers who run coding agents and tools such as OpenCode, Cline, OpenClaw, Hermes or the OpenAI SDK for many hours a day. The model supports text, images and tool calls, and paid users can also select an optional route with a configurable server side target that is initially backed by Flash. The operator runs its own model serving stack on dedicated leased capacity rather than reselling another provider's API. Prompt and response bodies are not stored or used for training; only operational metadata such as model, token counts, status and latency is kept. The Free plan costs $0 and allows 15 requests a week, resetting each Monday at 00:00 UTC, with a 128K context window and no card required. Builder is $19 per month with no per token billing, 128K context, approximately 200 million tokens a day and up to 5x concurrency. Pro is $39 per month with 256K context, approximately 750 million tokens a day and up to 10x concurrency. Paid plans renew monthly and can be cancelled at any time, and an enterprise offering covers dedicated GPU clusters, custom models and private inference. The catalogue is deliberately narrow: a single model family is offered, which the company says keeps serving efficient enough to sustain flat pricing.