Refs #260. This covers the denoise half. The issue's audio-output half (io.lotus.audio_output_state) is in the #310 work, waiting on the fork's lotus.24 package.
Problem
The fork reports what noise suppression actually runs as io.lotus.denoise_state { active, model, error? }. The host never listened ("Unknown or unsupported from-widget action"), so when ML didn't attach, the call silently published the raw mic while Settings still said ML was on.
The only listener left, LotusDenoiseFeature, waited for a {type: 'lotus-denoise-status'} postMessage that nothing sends any more. It also accepted that message from any frame, with no source or origin check, and showed its text as a system toast.
Change
New useDenoiseNotices(embed), mounted next to useScreenshareNotices. When the user has ML selected:
active: false → "ML noise suppression couldn't start, so your microphone is sending unfiltered audio. You can pick another model in Settings → Calls."
A different model → "DeepFilterNet 3 couldn't start, so noise suppression is using RNNoise instead."
LotusDenoiseFeature is deleted.
Verified with a real failure: on lotus the dev server 404s the worklets (#293), with ML / RNNoise selected and joining the Voice Lounge.
Before: no toast, plus 1 "host did not ack io.lotus.denoise_state".
After: the toast above, and 0 no-ack logs.
Unit tests for the texts. tsc clean · unit tests pass · ESLint clean.
Refs #260. This covers the denoise half. The issue's audio-output half (`io.lotus.audio_output_state`) is in the #310 work, waiting on the fork's lotus.24 package.
**Problem**
- The fork reports what noise suppression actually runs as `io.lotus.denoise_state { active, model, error? }`. The host never listened ("Unknown or unsupported from-widget action"), so when ML didn't attach, the call silently published the **raw mic** while Settings still said ML was on.
- The only listener left, `LotusDenoiseFeature`, waited for a `{type: 'lotus-denoise-status'}` postMessage that nothing sends any more. It also accepted that message from **any frame**, with no source or origin check, and showed its text as a system toast.
**Change**
- New `useDenoiseNotices(embed)`, mounted next to `useScreenshareNotices`. When the user has ML selected:
- `active: false` → "ML noise suppression couldn't start, so your microphone is sending unfiltered audio. You can pick another model in Settings → Calls."
- A different `model` → "DeepFilterNet 3 couldn't start, so noise suppression is using RNNoise instead."
- `LotusDenoiseFeature` is deleted.
**Verified** with a real failure: on `lotus` the dev server 404s the worklets (#293), with ML / RNNoise selected and joining the Voice Lounge.
- **Before:** no toast, plus 1 "host did not ack io.lotus.denoise_state".
- **After:** the toast above, and 0 no-ack logs.
Unit tests for the texts. `tsc` clean · unit tests pass · ESLint clean.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_017PCb59uY3UKqDqt7B6hzd5
The fork reports what actually runs as io.lotus.denoise_state, but the
host never listened ("unsupported action"), so a failed ML denoiser
silently published the raw mic. Handle it: toast when ML did not start
or fell back to another model. Remove LotusDenoiseFeature, which waited
for a postMessage nothing sends any more and accepted it from any frame.
Refs #260
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017PCb59uY3UKqDqt7B6hzd5
jared
merged commit 9c3fdfff39 into lotus2026-10-18 10:53:40 -04:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Refs #260. This covers the denoise half. The issue's audio-output half (
io.lotus.audio_output_state) is in the #310 work, waiting on the fork's lotus.24 package.Problem
io.lotus.denoise_state { active, model, error? }. The host never listened ("Unknown or unsupported from-widget action"), so when ML didn't attach, the call silently published the raw mic while Settings still said ML was on.LotusDenoiseFeature, waited for a{type: 'lotus-denoise-status'}postMessage that nothing sends any more. It also accepted that message from any frame, with no source or origin check, and showed its text as a system toast.Change
useDenoiseNotices(embed), mounted next touseScreenshareNotices. When the user has ML selected:active: false→ "ML noise suppression couldn't start, so your microphone is sending unfiltered audio. You can pick another model in Settings → Calls."model→ "DeepFilterNet 3 couldn't start, so noise suppression is using RNNoise instead."LotusDenoiseFeatureis deleted.Verified with a real failure: on
lotusthe dev server 404s the worklets (#293), with ML / RNNoise selected and joining the Voice Lounge.Unit tests for the texts.
tscclean · unit tests pass · ESLint clean.🤖 Generated with Claude Code
https://claude.ai/code/session_017PCb59uY3UKqDqt7B6hzd5
The fork reports what actually runs as io.lotus.denoise_state, but the host never listened ("unsupported action"), so a failed ML denoiser silently published the raw mic. Handle it: toast when ML did not start or fell back to another model. Remove LotusDenoiseFeature, which waited for a postMessage nothing sends any more and accepted it from any frame. Refs #260 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017PCb59uY3UKqDqt7B6hzd5