QUICK SETUP DOCS GUIDES

Houtini: Your AI's Local Superpower

Improve your coding and creation capabilities by offloading the grunt work, saving thread length in Claude Desktop or API calls in your IDE.

Use our powerful prompt library or let Claude take the lead with custom prompting. Houtini connects your AI workflow to a powerful, private LLM on your own machine or your company network. Go from idea to production in hours, not months.

OK, LM let's build a

Quick Setup

Install Houtini LM with npm and add this configuration to your mcp json configuration. If you're not sure and you need a hand, no problem - head to our comprehensive setup guide and learn how to install node and all other prerequisites in Windows or MacOS

1

Install the package

npm install -g @houtini/lm@latest
2

Add to Claude Desktop config

{
  "mcpServers": {
    "houtini-lm": {
      "command": "npx",
      "args": ["@houtini/lm@latest"],
      "env": {
        "LM_STUDIO_URL": "ws://127.0.0.1:1234",
        "LLM_MCP_ALLOWED_DIRS": "C:/your-projects,C:/your-code"
      }
    }
  }
}

Update the LLM_MCP_ALLOWED_DIRS path to match your project directories

Requires

Install these essential tools for your local AI workflow

LM Studio

Your local LLM powerhouse

Install from lmstudio.ai

Desktop Commander

File operations & system integration

Install from desktopcommander.app
How it works
Your AI's local superpower in action
You
Claude Desktop
Your interface
Claude Code
Your IDE
Cursor IDE
Your editor
Houtini
{ }
AI
Your AI Superpower
Intelligent routing
Free
LM Studio
Grunt work
Unlimited
Premium
Claude API
Strategic thinking
Preserved budget
Unlimited Generation
No token limits, no usage caps
30+
Expert Functions and growing
Use our list of battle-tested prompts

Everything you need to build faster

30+ expert functions, local processing, and seamless integration with your existing workflow.

30+ Expert Functions

Use our list of battle-tested prompts for code analysis, security audits, documentation generation, and WordPress development. Each function delivers professional-quality output with hundreds of hours of refinement built in.

Local Processing

Avoid hitting those pesky thread length limits in Claude Desktop. Run unlimited generations on your own hardware through LM Studio. Avoid API rate limits with no usage cap, and no more rewriting the same prompt over and over for repetitive tasks.

Seamless Integration

Works natively with Claude Desktop and Claude Code through MCP. Preserve your Claude context for strategic thinking whilst handling routine tasks locally. Setup takes under 5 minutes.

Built for People Who Build

Professional Developers

Generate scaffolding, run security audits, create documentation - all with production-ready code that follows best practices. Unlimited local processing, extensible architecture, and expert-quality output that passes code reviews.

Vibe Coders

Transform inspiration into working code instantly. Analyze what works elsewhere and adapt it to your creative vision. No boilerplate, no configuration hell, no API limits - just pure creative flow from idea to deployed application.

Agencies & Freelancers

Deliver enterprise-grade solutions at startup speed. While competitors wrestle with API budgets and token limits, you'll be shipping production-ready code with unlimited iterations and professional confidence.