Qiaomu LLM MCP — AI skill for Claude Code
把多模型 Provider、本地密钥和 HeavySkill 讨论统一成 MCP 网关 Local MCP gateway for multi-provider LLM routing, secrets, and HeavySkill discussions.
How to install Qiaomu LLM MCP
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open joeseesun/qiaomu-llm-mcp and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Qiaomu LLM MCP does
把多模型 Provider、本地密钥和 HeavySkill 讨论统一成 MCP 网关 | Local MCP gateway for multi-provider LLM routing, secrets, and HeavySkill discussions.
Alternatives in AI
- Loushang — AI-native agent harness for coding workflows by python: multi-model LLM orchestration, stateful sessions, tool 1.3k ★
- My Free Code — Open-source multi-provider AI gateway for Claude Code and other coding agents, with model routing, streaming 633 ★
- Ultimate MCP Server — Comprehensive MCP server exposing dozens of capabilities to AI agents: multi-provider LLM delegation, browser 146 ★
README
qiaomu-llm-mcp
把多个 LLM Provider、API Key、模型路由和多模型讨论统一成一个本地 MCP 网关。
A local MCP gateway that turns many LLM providers, API keys, routing rules, and multi-model discussions into one tool surface.
**中文** | [English](#english)
适合这些场景:
- 你同时使用 Z.ai Coding Plan、DeepSeek、OpenAI-compatible API、Anthropic-compatible API、Claude Code CLI 等多套模型。
- 你不想把 API Key 写进提示词、脚本或 Codex 配置里。
- 你希望在 Codex / Claude / 其他 MCP 客户端里用同一组工具完成:单模型调用、多模型对比、流水线、HeavySkill 式多轮讨论。
- 你需要让最强的主控模型主持讨论,而不是把全部判断交给某一个 Provider。
本项目是本地优先工具:Provider 注册表保存在 `~/.config/qiaomu-llm/registry.json`,密钥默认放在 macOS Keychain,MCP server 只暴露“密钥是否存在”,不会返回明文 Key。
你能得到什么
安装后,MCP 客户端会看到这些工具:
| 工具 | 作用 |
|---|---|
qllm_list_providers |
列出本地配置的 Provider,并安全显示密钥状态 |
qllm_show_provider |
查看单个 Provider 的非敏感配置 |
qllm_list_models |
查看本地配置或远程 /models 返回的模型 |
qllm_route_task |
不花 token,根据任务类型选择 Provider / Model |
qllm_chat |
调用单个 Provider |
qllm_compare |
用同一个问题对比多个 Provider |
qllm_pipeline |
串联多个模型步骤完成任务 |
qllm_heavy_discuss |
多视角、多模型讨论,可由模型或 Codex 主持 |
qllm_heavy_render |
把 Codex 主持后的讨论渲染成 Markdown / HTML 报告 |
qllm_claude_code_* |
通过 MCP 安全调用本机 Claude Code CLI |
推荐工作流
用户问题
|
+-- qllm_route_task 选择模型
|
+-- qllm_chat / qllm_compare / qllm_pipeline
|
+-- 复杂议题: qllm_heavy_discuss
|
+-- 多模型独立给观点
+-- Codex 或指定模型主持
+-- qllm_heavy_render 输出 Markdown / HTML
这套设计的核心不是“多叫几个模型凑热闹”,而是让不同模型承担不同角色:快速模型做初稿,强推理模型做审查,写作模型做表达,Codex 作为当前工作区里的主持人整合上下文和行动方案。
快速开始
1. 准备环境
要求:
- macOS 或 Linux
- Python 3.10+
- 一个支持 MCP 的客户端,例如 Codex、Claude Desktop、Claude Code、Cursor 等
- 至少一个 LLM Provider 的 API Key
克隆项目:
git clone https://github.com/joeseesun/qiaomu-llm-mcp.git
cd qiaomu-llm-mcp
如果你还没有 GitHub 版本,也可以先在本地目录安装:
cd /path/to/qiaomu-llm-mcp
2. 安装 Python 依赖
python3 -m venv .venv
.venv/bin/python -m pip install -U pip
.venv/bin/python -m pip install -e .
确认可执行文件存在:
test -x .v
Related Skills
Req LLM Gateway
OpenAI-compatible LLM proxy with telemetry and multi-provider routing
Agent Switch
A free, lightweight AI gateway manager that lets Codex and Claude clients share a switchable, observable local
Gremlord
Run Claude Code on any model, with a budget — multi-provider router, LLM-triaged tier routing, spend tracking
Token Station
Local routing control plane for AI agents and LLM providers, with a loopback-only gateway, smart and quota-awa
Evolve Router
A local-first LLM routing gateway that thinks before it routes — a decision model (Jev / laya) picks the most
Cab
One local gateway for every coding agent. Cost-aware LLM model routing for Claude Code, Codex, OpenCode, and m
Related Agents
LLM Orchestrator
Use this agent for LLM integration work — prompt engineering, multi-provider abstraction (OpenRouter, Gemini,
Jordylab DevOps
JordyLab production DevOps: the OVH VPS + k3s deployment, Tailscale access, CI/CD deploys and rollbacks, SOPS
Local Worker
Coding worker on a local model (Ollama, LM Studio, vLLM) served by the multi-model gateway's upstream. Use whe