zero-token

永久免 token 费的 openclaw， ChatGPT、Claude、Gemini、DeepSeek、千问、豆包、Kimi、智谱清言均可免费；三个重磅 Skill 大幅提升桌面执行能力；无需安装，开箱即用；支持微信、钉钉等多端接入控制。兼容所有 OpenClaw 技能，并提供更强的 Skill 扩展。交流社区：钉钉群165930016082，需更快更稳定能接入Claude code的版本私

A lightweight AI gateway that exposes an OpenAI-compatible interface with full Tools support, using web AI sessions instead of API tokens.

对网页deepseek的封装，实现免费API调用。支持 OpenAI completions / responses socket。可工具调用，可接入codex、copilot。轻量自用

ONGRID — Ontology-Native Graph · Retrieval · Inference · Decision. 3D knowledge graph + zero-token ZTC answer verification. by VIDRAFT.

Local, zero-token intent routing for chat, tooling, coding, and planning prompts.

🛡️ Zero-token model-independent verification layer for AI coding agents

Run DeepSeek Harness on desktop with zero token costs—a free, open-source GUI for macOS and Windows.

Darwin-27B-ZTC-v2 — #1 of 102 on the System One Mosaic Benchmark (S1MB). A typed-decision judge that scores decisions in one forward pass, zero generated tokens (ZTC). by VIDRAFT.

ZTC (Zero-Token Confidence) — verify whether an LLM answer can be trusted in one forward pass, zero generated tokens. ZTC-Judge + Darwin-ZTC family. #1 on S1MB. by VIDRAFT.

Unofficial standalone WorkBuddy daily credits automation for macOS

Advanced zero-token LLM decision engine in Rust synthesizing architectural breakthroughs from open-source Jev alternatives.

Zero-token query lane for WeCom: type "@bot <command>" in a group and a local read-only script answers in the same message. No LLM in the loop, no context replayed, millisecond replies from a background snapshot cache. Allow-listed commands, safe to hand to colleagues.

Zero-token on-demand translation for the DeepSeek Harness Web GUI — hover a block or drag-select text to translate in place; nothing auto-translates, no LLM calls, text never leaves your machine. ｜ 零 token 按需翻译：悬停或框选才翻，绝不自动翻页，文本不出本机。

DSH 运行时监察：工具调用 / 事件循环 / 错误风暴实时体检 + 按调用强制停止 + 右下角红绿灯挂件（0 token）

AI is promising to streamline process! Can AI help us with seamless integrations too?

0TA - Zero Token Architecture: Zero tokens. Infinite thoughts. Stop prompting. Start thinking.

Zero-token checks for Claude Code setups (stale memory, broken hooks, retired references, reachability), plus a cache keepalive hook and ntfy notifications.

Decision Index: reproduce the Jev decision-model benchmark suite locally or as one HF Job

Zero-token task router for AI coding agents — classifies requests locally without burning LLM tokens. 16 archetypes, 4 complexity tiers, in-context classification.

Zero-token, server-less usage & rating tracker for Claude Code skills and subagents — counts live in plain markdown in your repo.