WebCaptioner went offline in late 2023. For years it was the free-and-friendly tool that small churches, classrooms, low-budget events, and accessibility users reached for when they needed live captions for in-person speech. When the site went down — citing sustainability — a sizable community was left with no good free replacement.
LiveCaptionIt isn’t a clone of WebCaptioner. The use cases overlap but the architecture is fundamentally different. Here’s the honest comparison.
What WebCaptioner did
- Browser-based, no install
- Free, no signup
- Used Chrome’s Web Speech API (server-routed to Google’s speech recognition)
- Output captions to a full-screen view or a customizable on-stage display
- Popular with houses of worship, conferences, classroom AT setups
The architecture had one big downside: Web Speech API requires a server round trip. Audio was streamed from your browser to Google’s recognition service. Free but not private.
What LiveCaptionIt does
- Browser-based, no install
- Free, no signup
- Runs Whisper locally via WebGPU + transformers.js
- Output to inline or floating Picture-in-Picture window
- Same use cases — churches, classrooms, accessibility, events
The key difference: audio never leaves your device. Whisper does the recognition in your browser using your GPU. No server round trip. No network dependency once the model is cached.
Feature-by-feature
| Feature | WebCaptioner (RIP) | LiveCaptionIt |
|---|---|---|
| Recognition engine | Google Web Speech API (server) | Whisper (local) |
| Privacy | Audio sent to Google | Audio never leaves device |
| Languages | English + ~30 others | 99 (Whisper) |
| Offline | No (server required) | Yes (after first model load) |
| Floating window | No | Yes (Document PiP) |
| Source toggle | Mic only | Tab + Mic |
| Caption styling | Basic | Themes + size + position + opacity |
| Custom vocabulary | No | Yes (Whisper initial_prompt) |
| Transcript download | Yes (.txt) | Yes (.txt + .vtt + .srt) |
| Session history | No | Last 20 in browser |
| Shareable link | No | Yes (gzipped in URL) |
| Open source | Yes (archived) | Yes (active) |
Where WebCaptioner did things better
Being honest:
- Display flexibility. WebCaptioner’s “stage display” mode was optimized for projector setups — large fonts, customizable bg/fg, output via a secondary monitor. LiveCaptionIt’s floating PiP works for desktop use but isn’t designed for stage projection.
- Lower hardware requirements. WebCaptioner just needed audio capture and an internet connection. LiveCaptionIt needs WebGPU and ~75 MB of model download — meaning older laptops or Chromebooks-without-GPU are slower.
If you specifically need a projector-style display for in-person events and don’t care about audio leaving your venue, the WebCaptioner archives may still partially work — some forks reportedly maintain limited functionality.
Where LiveCaptionIt is strictly better
For everything else:
- Privacy is non-negotiable. Hospitals, therapists, lawyers, school counselors — anyone whose audio shouldn’t go to a third party — can use LiveCaptionIt without compromising compliance.
- Floating window for one-on-one accessibility. A hearing-impaired user can put LiveCaptionIt’s PiP window above any meeting, video, or conversation app. WebCaptioner could only run as a full-screen display.
- Better language coverage. Whisper supports 99 languages out of the box — much wider than WebCaptioner’s Google-API-backed list.
- Custom vocabulary. Teach the model your proper nouns, technical terms, jargon. Pre-prompt support means it knows “kubectl” is a word and not “cube control.”
- True offline. Cache the model once, use forever without network.
How to migrate
If you used WebCaptioner before its shutdown:
- Open LiveCaptionIt in Chrome / Edge / Brave 116+
- Click Start, pick your audio source (microphone for in-person events, tab for online meetings/videos)
- Open the Caption Style panel and adjust font size + position to match your old WebCaptioner setup
- For an “always-visible” display similar to WebCaptioner’s stage mode, pop out the floating PiP window and resize edge-to-edge horizontally for a subtitle-bar look
For projector display, drag the PiP window onto your secondary monitor and fullscreen it (most desktop OSes support fullscreening individual windows).
Closing thought
WebCaptioner’s shutdown left a real gap. We’re not the same tool, but we’re trying to serve the same community better — free, private, multilingual, offline-capable. If you used WebCaptioner and have specific workflows that don’t translate, I’d genuinely like to hear about them. Issues open on GitHub.