Adjusts loudness so soft and loud speech sound the same.
Test input
Speak and pause to see your levels. Auto-stops after 3 seconds,
then you can replay the clip.
Listening… speak now.
Audio output
Select device
Where audio (TTS, dubbed clips) is played.
Not supported in this browser.
Volume control
Master playback volume for every audio element on the page.
Minimum 10% — use Mute below to silence everything.
80%
Mute audio
Silences every audio element on the page (clip playback, Suite,
Master stream, and the test chime). Volume is preserved when
you un-mute.
Stereo / mono
Force a single-channel mix or keep stereo separation.
Spatial audio
Position different speakers in 3D space (where supported).
Test output
Play a test tone and watch the output level to check your audio.
Output level
Subtitle
Show subtitle
Show the floating subtitle overlay as you speak or listen.
Caption list (with audio)
Show a scrollable list of captions, one per line, each with a play button for its audio.
Show as
Show the original words, the translation, or both.
Translated to
Language to re-translate the subtitle into on the fly.
Position
Where the subtitle overlay sits on screen.
Text size
Size of the subtitle text.
Background opacity
Opacity of the subtitle overlay background.
Translation
Show player
Reveal or hide the floating translation player.
Listen as
Hear the translated dub or the speaker's original audio.
Real voice (Original)
In Original mode, hear the speaker's actual voice instead of a synthetic read-back.
Falls back to the synthetic voice when it isn't available.
Dub language
Which language to hear the TTS dub in.
Live translation
Translate and broadcast your speech in real time, voice included.
Translate to
Language to translate your speech into.
Voice
Voice used for spoken playback. Updates with the target language.
English voice
Voice used when dubbing into English.
French voice
Voice used when dubbing into French.
German voice
Voice used when dubbing into German.
Position
Where the translation player sits on screen.
Text size
How large the translation player text appears.
Background opacity
How solid the player surface looks — low values give a
transparent, subtitle-style look.
Appearance
Theme
Light, dark, or match your system.
Layout
Free: drag the caption and player anywhere. Fixed: docks
them as top and bottom bars.
Bar width
In Fixed layout, span the full width or keep side margins.
Descriptions
Show the explanatory text under each setting.
Section headers
Show the heading above each settings group.
Accent color
Highlight color for active controls and selections.
Reduced motion
Minimize animations and transitions.
Interface language
Language of the settings panel.
Engines
Choose where each capability runs. Preview only — these
controls remember your choice but do not change behavior yet.
Defaults match the current app, so nothing regresses.
Speech recognition
Off · App = browser Web Speech · WebGPU = local Whisper in-browser
(offline, private) · Cloud = your Whisper server.
Text-to-speech
Smart = adapts Cloud vs offline per phrase (tune the mix below) ·
Auto = Edge-TTS, falls back to the browser voice if unreachable ·
None = off · App = in-browser voices (offline) · Cloud = Edge server only.
Smart mix — how much each factor matters
Speed
50
Reliability
75
Quality
80
Privacy
20
Translation
Off, or your MarianMT server. (No offline browser option.)
Correction
Clean up recognized text (capitalization, punctuation, typos).
App = instant offline fixes; Cloud = server /correct.
Enhancements
Glossary fix
When Correction = baai: post-edit the translation against the local textile
termbase (instant, offline) and underline the words it corrected.
Live partial captions
Sync in-progress (interim) captions to participants in real time, shown with a
provisional underline until the phrase is final. Original language only.
Translate live partials
Also translate the in-progress (interim) caption into your read-language, shown greyed and
tentative until the final lands. Requires Live partial captions. Beta.
Show my translation (verify)
As the speaker, show a live translation of your own speech below the source line, so you can
verify the translation as you talk. Needs a translated/both view. Beta.
Incremental dub (streaming)
Start dubbing each clause as you speak, instead of waiting for the whole sentence — the final
still plays complete (listeners de-dup). As the speaker, commits stabilized clauses. Beta.
Correction highlight
Flash only the words that change when a better result lands.
Caption word colors
Draft / verified / corrected color states.
Live metrics
Off = hidden. Normal = words, commits, finalize lag, % rewritten.
Advanced = all metrics + charts + connection quality.
Live metrics — legend
We need a new deployment strategy.
Nous avons besoin d'une nouvelle stratégie de déploiement.