I built Voicebun as a pet project because I was frustrated with how locked-down and rigid voice agents still are—many tools are either closed platforms, lack deep customization, or charge a lot just to test ideas. I wanted a way to quickly spin up AI voice agents that I could fully control—from the voice itself to the logic behind every word.
Voicebun is an open-source voice agent builder. You can:
- Clone voices and generate audio using your own scripts and prompts - Dynamically change prompts based on user input - Add custom logic through functions and webhooks (e.g. hit an API, change the flow, log events) - One click deploy
Live product: https://www.voicebun.com Code: https://github.com/phombal/voicebun
I made this during a period of intense prototyping and wanted something more modular that developers could hack on. It’s built for speed, iteration, and integration into real workflows, whether it’s a support bot, creative narrator, or something weird.
Would love to hear your feedback, use cases, or suggestions for what to build next!