Knights Address · Open-source AI speech coach
Practice out loud! Get coaching that actually helps you deliver.
Record a practice speech with your camera and mic. In under a minute you'll see your pace, pauses, filler words, clarity, and how well you fit your audience, with three specific things to work on next.
No sign-up. Your guest profile stays on this browser.
How it works
- Step 1
Record
Pick your audience, then speak for 10 seconds to 5 minutes with your camera and mic.
- Step 2
Measure
Whisper transcribes your speech and we measure pace, pauses, and filler words.
- Step 3
Coach
An open-source Llama model reviews clarity, structure, and audience fit.
- Step 4
Improve
Get three specific fixes, quotes from your own words, and a practice drill.
Graded for your audience
Speaking to kids, teens, or seniors? Pick your audience and your pace targets, vocabulary level, and coaching adapt to what works for them.
What we capture, and where it goes
- Video stays on your device. It's used for your self-view and an instant replay, and it's never uploaded.
- Audio is sent for transcription and not stored. It goes to an open-source Whisper model on Hugging Face and is discarded afterward.
- Transcripts and scores are stored in your private guest profile so you can track progress. Delete them anytime.