HN Showcase
Back to browse
GitHub Repository

An AI-infused teleprompter that runs entirely in the browser

24 starsJavaScript

Voice-tracked teleprompter using on-device ASR in the browser

by lbaune·Mar 15, 2026·6 points·1 comment
Visit ProjectView on HN

AI Analysis

●●SolidNiche GemBig Brain

Voice-tracked scrolling runs entirely in-browser using Moonshine ONNX, no backend needed.

Strengths
  • •Moonshine ONNX model runs in Web Workers, keeping UI thread free during transcription.
  • •Fuzzy matching recovers position if you skip lines or ad-lib sentences.
  • •Zero data leaves the browser tab, ensuring privacy for sensitive scripts.
Weaknesses
  • •Browser-based ASR accuracy varies significantly across different devices and microphones.
  • •Niche use case limits adoption compared to general-purpose recording tools.
Category
Productivity
Target Audience

Video creators, presenters, streamers

Similar To

Descript · Teleprompter Premium · CapCut

Post Description

I built a teleprompter that scrolls based on your voice instead of a timer.

Paste a script, press record, and it highlights the current word as you speak. If you pause it waits; if you skip lines it finds its place again.

Everything runs entirely in the browser — speech recognition (Moonshine ONNX), VAD, and fuzzy script matching.

Demo: https://larsbaunwall.github.io/promptme-ai

Most of the project was initially built using Perplexity Computer, which made for an interesting agentic coding workflow.

Curious what people think about the script alignment approach.

Similar Projects