Sahil Manocha

Hey folks We built a dashboard that shows running cost and supports multiple speech-to-speech models (Realtime API, Outspeed hosted MiniCPM-o). It is based on OpenAI's realtime console and agents UI....

fast MiniCPM-o deployed endpoint (3x faster inference)

2

We have optimized MiniCPM-o and created fast MiniCPM-o with ~3x inference speed for voice-to-voice chat. You can try it out here: https://demo.outspeed.com/