clauden: run Claude Code on free NVIDIA-hosted models via a LiteLLM proxy
gmaggiotti · x · 2026-10-13
NVIDIA's inference API currently hosts Kimi, DeepSeek, Llama, Qwen and more for free. Developer gmaggiotti open-sourced clauden, a CLI wrapper that points Claude Code at those free models instead of Anthropic's.
How it works: Claude Code speaks the Anthropic API while NVIDIA NIM speaks OpenAI's, so clauden spins up a local LiteLLM proxy (127.0.0.1:4000) that translates between the two, maps every Claude model tier to your chosen NVIDIA model, and launches Claude Code against it.
Usage: clauden kimi-k3 — same agent, same tools, free models.
More from coding & agent
- Stop building coding loops with Jev — it's a general-purpose semantic judgment function — SpreadUsual4084 · 2026-10-13
- 'Set up in 5 minutes or it's not a personal agent': ranking debate ignites — sujingshen · 2026-10-13
- Developer builds M5Stack hardware to enable iPhone Use around iOS sandboxing — yangyi · 2026-10-13
- Codex plus Blender: using an AI coding agent for planning and iteration — CSProfKGD · 2026-10-13
- REA hits 120k stars: when AI can reverse-engineer any software, what's left? — Arindam_1729 · 2026-10-13
- Open-source Claude people-search skill delivers verified leads at $0.029 each — ramagetime · 2026-10-13