Work in Progress: This guide is being actively developed.

Local AI Assistant

Run AI on your own machine. Self-hosted and open source.

Get Started Free

Why Run AI Locally?

🖥️

Self-Hosted

Run on your own hardware. You manage the setup and configuration.

🌐

Works Offline

With local models, ClawdBot can work without internet.

⚡

Lower Latency

No network round-trips can mean faster responses for some tasks.

💰

No API Costs

With local models, there are no per-request charges. Hardware is your main cost.

How Local AI Works

ClawdBot Runs on Your Computer

Unlike Siri, Alexa, or ChatGPT, ClawdBot runs entirely on your Mac, Windows, or Linux machine. No cloud connection required for core functionality.

Choose Your AI Backend

You decide where AI processing happens:

  • Local models (Ollama) - Runs offline on your hardware
  • Cloud APIs (Anthropic/OpenAI) - Better quality, requires internet
  • Hybrid - Use local for simple tasks, cloud for complex ones

Configure Access

Configure what ClawdBot can access on your system - files, emails, calendar, etc.

Local AI Options

Fully Local
🦙

Ollama + Local Models

Run Llama, Mistral, and other open models directly on your hardware.

  • Runs offline
  • No API costs
  • Requires decent hardware (16GB+ RAM)
Get Ollama
🧠

Anthropic Claude API

High quality responses via cloud API.

  • High quality responses
  • Cloud-based API
  • Pay-per-use pricing
Get API Key
🤖

OpenAI GPT API

Alternative cloud option with familiar GPT models.

  • GPT-4 quality
  • Wide compatibility
  • Pay-per-use pricing
Get API Key

Hardware for Local AI

Want to run AI completely locally? Here's what you need:

Minimum Requirements

  • RAM: 16GB minimum (for 7B parameter models)
  • Storage: 10-50GB for models
  • CPU: Modern multi-core (M1+ Mac or recent Intel/AMD)

Recommended for Best Performance

  • RAM: 32GB+ (for larger models)
  • GPU: NVIDIA with 8GB+ VRAM (for fast inference)
  • Storage: SSD with 100GB+ free

Great Local AI Setups

  • Apple Silicon Mac: M1/M2/M3 with unified memory - excellent for local AI
  • Gaming PC: RTX 3060+ GPU runs models very fast
  • Mini PC: Intel NUC with 32GB RAM for always-on assistant

Set Up Your Local AI

# 1. Install ClawdBot
curl -fsSL https://clawd.bot/install.sh | bash

# 2. Install Ollama for local models (optional)
curl -fsSL https://ollama.ai/install.sh | bash
ollama pull llama2  # or mistral, codellama, etc.

# 3. Configure ClawdBot
clawdbot onboard --install-daemon

# 4. Start using your local AI assistant!
Full Setup Guide