Skip to content

INSTALLATION.md

Latest commit

 

History

History
198 lines (137 loc) · 3.37 KB

INSTALLATION.md

File metadata and controls

198 lines (137 loc) · 3.37 KB
Raw

CrawlLama - Installation & Start

Navigation: Home | Docs | Quickstart | LangGraph

Quick Installation

Windows

  1. Run setup:
setup.bat
  1. Start Ollama (separate terminal):
ollama serve

Note: The first dependency installation (handled by uv from pyproject.toml) may take 5–10 minutes or longer for packages like torch and sentence-transformers. Wait until installation completes.

  1. Download Model:
ollama pull qwen3:8b
  1. Start CrawlLama:
run.bat

Linux/macOS

  1. Run setup:
chmod +x setup.sh run.sh
./setup.sh
  1. Start Ollama (separate terminal):
ollama serve
  1. Download Model:
ollama pull qwen3:8b
  1. Start CrawlLama:
./run.sh

What setup.bat / setup.sh Does

  1. Installs/uses uv (which manages the pinned Python interpreter and .venv)
  2. Creates the virtual environment (.venv)
  3. Installs dependencies from pyproject.toml
  4. Creates directories (data/, logs/, plugins/)
  5. Copies .env.example → .env and config/config.json.example → config.json
  6. Verifies Ollama installation

Using run.bat / run.sh

Always use the run scripts to activate the virtual environment automatically.

Examples:

# Interactive mode
run.bat

# Direct question
run.bat "What is Python?"

# Options
run.bat --no-web "Offline question"
run.bat --debug "Debug mode"
run.bat --stats
run.bat --clear-cache

Manual venv Activation

Windows:

.venv\Scripts\activate
python main.py

Linux/macOS:

source .venv/bin/activate
python main.py

Directory Structure After Installation

crawllama/
├── .venv/ # Virtual environment
├── data/
│ ├── cache/ # Web cache
│ ├── embeddings/ # ChromaDB
│ └── history/ # Chat history
├── logs/
│ └── app.log # Logs
├── setup.bat/setup.sh # Setup scripts
├── run.bat/run.sh # Start scripts
└── ...

Common Issues | Issue | Solution |

| -------------------- | ------------------------------------------------- | | venv not found | Run setup.bat or ./setup.sh again | | Ollama not running | Start Ollama: ollama serve in separate terminal | | Model not found | Download: ollama pull qwen3:8b | | Missing dependencies | Run setup script again |

Alternative Models

# Default (recommended)
ollama pull qwen3:8b

# Strong reasoning
ollama pull deepseek-r1:8b

# Faster, smaller
ollama pull qwen3:4b

# Larger, more accurate
ollama pull llama3:7b
ollama pull mistral:7b

# Very high performance, more RAM
ollama pull phi3:14b

Disk Space Note: After setup + models, expect 1–2 GB minimum; larger models may require 6–20+ GB depending on format.

Test Installation

# Show system stats
run.bat --stats

Expected output:

{
 "tools_available": 3,
 "web_enabled": true,
 "model": "qwen3:8b",
 "cache": {...}
}

Support