OneCielAI/claude-any

Claude Any is a compatibility launcher and router that lets Claude Code run unchanged while using non‑Anthropic or lower‑cost LLMs (NVIDIA NIM, Ollama/Ollama Cloud, DeepSeek, LM Studio, vLLM, local Ollama, OpenCode, etc.). It’s aimed at operators, developers, and teams who need continuity when Anthropic access is limited or want to run Claude Code workflows on free/hosted/self‑hosted models. Key features: - Pre‑launch provider/model selection (URL, API key, streaming options) while preserving Claude Code tooling and arguments. - Plan Mode preservation and optional long‑context Advisor model for reviews and feedback-in-transcript. - Router with local browser chat (/ca/web/chat), SSE/channel endpoints (/ca/channel/*) for realtime agent bridging, and RPM pacing to smooth free‑model limits. - Operator controls: per-user router ports, observability/event views, live LLM settings, safe external debug toggle, and model capability overrides (ultracode/workflows flags). - Compatibility improvements and provider presets (DeepSeek, Ollama Cloud, NVIDIA hosted), plus better transcript normalization and Plan Mode loop recovery.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.