Hi everyone,

I wanted to share a project I’ve been working on: a voice-operated conversational AI with the personality of HAL 9000 from 2001: A Space Odyssey. It’s all built in Python and runs 100% locally.

What My Project Does

Voice Control: Uses voice input from your microphone to converse with the AI Local LLM: Uses local models with Ollama for entirely offline LLM inference/responses HAL Personality: Features a custom StyleTTS2 voice model fine-tuned with voice data from 2001: A Space Odyssey. MCP Tooling Support: Supports Weather and Time MCP tools

Target Audience

This project is mainly for everyone (especially Space Odyssey fans) who want to play with a unique voice-controlled AI assistant. It’s also for enthusiasts who are interested in local AI, real-time speech transcription, and real-time voice synthesis.

Comparison

Many other AI voice assistants use generic voices, but the HAL 9000 voice used in this project was fine-tuned for over 30 hours on each line of dialogue from HAL 9000 in 2001: A Space Odyssey. While most other AI voice assistants rely solely on the LLM’s built-in knowledge base, this project supports MCP servers like weather, time, and web search Some other LLM assistants rely on cloud APIs, but HAL 9000 can run entirely offline

Try It

Check out the Github repo: https://github.com/tizerk/hal9000

It’s been a great learning experience working on HAL, and I’m hoping to add a lot more to it.

Feedback is highly appreciated. The code definitely isn’t the cleanest or most optimized, and I would love to hear some solutions.

Let me know what you think!