Uploaded March 2025 | Updated September 2026, 2 weeks ago
I will turn the skeleton from my last project into a cohost. Openai Faster-whisper does speech to text. That gets piped to a locally hosted LLM using ollama. The output uses Piper TTS. Cap for testing, as is seen here, I can let a friend remotely "possess" Timmy to control his head pan tilt servos, see-through his eye (using H.264HLS streaming), and enter text commands into a flask python server better than red allowed by Piper.
Very simple straightforward project.
I will turn the skeleton from my last project into a cohost. Openai Faster-whisper does speech to text. That gets piped to a locally hosted LLM using ollama. The output uses Piper TTS. Cap for testing, as is seen here, I can let a friend remotely "possess" Timmy to control his head pan tilt servos, see-through his eye (using H.264HLS streaming), and enter text commands into a flask python server better than red allowed by Piper.
Very simple straightforward project.
