Nano LLM Proxy — AI skill for Claude Code
Tiny single-binary LLM gateway — pool multiple providers and API keys behind one OpenAI/Anthropic-compatible endpoint, with smart key rotation and an embedded admin GUI.
How to install Nano LLM Proxy
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open kmmuntasir/nano-llm-proxy and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Nano LLM Proxy does
Tiny single-binary LLM gateway — pool multiple providers and API keys behind one OpenAI/Anthropic-compatible endpoint, with smart key rotation and an embedded admin GUI.
Alternatives in AI
- Pixel Agents — Compressed Reference — VS Code extension with embedded React webview: pixel art office where AI agents (Claude Code terminals) are an 6.5k ★
- Embeddedskills — An open-source collection of embedded development and debugging skills for Claude Code, Copilot, TRAE, and oth 593 ★
- Claudexor — Multi-harness control plane for Claude Code, Codex, Cursor, and OpenCode: quota-aware rotation across multiple 425 ★
README
nano-llm-proxy
[](https://go.dev) [](https://www.gnu.org/licenses/gpl-3.0) [](https://github.com/kmmuntasir/nano-llm-proxy/actions/workflows/ci.yml) [](https://pkg.go.dev/github.com/kmmuntasir/nano-llm-proxy)
**A tiny, single-binary LLM gateway. Five direct Go dependencies, all pure Go — no cgo, no container runtime, no Postgres.** One static binary plus one SQLite file is the entire deployment.
[](https://github.com/kmmuntasir/nano-llm-proxy/releases/latest) [](LICENSE)
Pool several upstream providers and their API keys behind one endpoint that speaks **OpenAI chat**, **OpenAI Responses**, and **Anthropic Messages** — with health-tracking key rotation, an embedded admin GUI, and no runtime dependencies beyond a single SQLite file.
Think "personal API gateway for language models": point your scripts, IDE extensions, and CLI agents at one URL with one key, while the gateway rotates a pool of upstream keys, survives rate limits and flaky upstreams, and shows you what happened in a web UI.
**No Docker, no sidecars, no Postgres, no control plane.** One static binary plus one SQLite file is the entire deployment. `sudo ./deploy.sh` builds it and installs a hardened systemd unit. The whole point is that you can drop it on a cheap VPS, a NAS, or a laptop and have a working gateway in under a minute — the same job most gateways make you spin up a container and a database for.
Related Skills
Io
A tiny, self-contained LLM coding agent. One binary, 180+ providers, zero runtime dependencies.
Nadar Code
An autonomous, fully-local AI coding agent (Desktop IDE & CLI) powered by OpenRouter. It reads, writes, and ex
Hub Cc
Local control plane for Claude Code on Windows and macOS: switch LLM gateways in one click behind a fixed endp
Bebok
Desktop GUI Headless Local-first AI coding agent in Rust (axum, tokio, rmcp, portable-pty) behind HTTP + SSE,
Ruby Nano Bots
Ruby Implementation of Nano Bots: small, AI-powered bots that can be easily shared as a single file, designed
Vmr
A local-first, single-binary LLM router & Agent flight recorder. Automatically turns byte-faithful logs into s
Related Agents
Castari Proxy
Use Claude Agent SDK and Claude Code with other providers/models.
Grok Agent 01
ONLY job: find free or freemium AI token/API plans (OpenAI-compatible, Anthropic, Google, open routers, studen
Router Client
Owns the Anthropic-compatible router HTTP client. Use for /healthz, /v1/models, scrub, RouterError, errorName,