Collection of installation scripts for running various AI tools locally, along with helper scripts for common AI-related tasks. Tested on Ubuntu.

All installation scripts are located in the install-scripts/ directory.

Stable Diffusion UI running on port 8188

./install-scripts/comfyui-docker.sh

Local LLMs installation with external access enabled

./install-scripts/ollama.sh

Web interface for Ollama on port 47586

./install-scripts/openwebui-docker.sh

Helper scripts for common AI-related tasks are located in the helper-scripts/ directory.

• file-to-transcript.sh - Transcribe audio files using OpenAI or Groq
• image-ocr.sh - Extract text from images using OpenAI GPT-4 Vision
• translate-audio.sh - Record and translate audio using OpenAI/Groq

• Docker with NVIDIA Container Toolkit
• NVIDIA GPU
• Ubuntu (or similar Linux)

git clone [email protected]:vlazic/ai.git
cd ai

# Run what you need
./install-scripts/ollama.sh            # Install Ollama
./install-scripts/openwebui-docker.sh  # Setup Ollama web UI

• ./install-scripts/storage/ and ./install-scripts/safetensors-models/ are gitignored (except .gitkeep)
• All Docker containers use NVIDIA GPU
• Sample FRP client configuration is provided in frpc.ini.sample for remote access. Put it in /etc/frp/frpc.ini and run sudo systemctl restart frpc to apply changes.

• Add Ollama proxy server for remote access ParisNeo/ollama_proxy_server