openclaw-zero-token

Why we included this project

This fork of OpenClaw cuts LLM costs by a different route than most gateways: instead of buying API tokens, it logs into the official web UIs of providers like ChatGPT, Claude, DeepSeek, and Qwen once through a browser, then serves them through a single interface you can call without per-request billing. It also routes requests through messaging channels like Telegram and Discord, and can inject local tool calls such as web search and file access into the web models that support them. Teams already running OpenClaw or similar agent gateways will find the zero-token fork a straightforward way to trim API spend while keeping one provider layer, though the browser-session approach is worth weighing against the stability and support you get from paid APIs.