Spaces:
Running
Running
German-Russian goes through English
Browse files- index.html +1 -1
index.html
CHANGED
|
@@ -89,7 +89,7 @@ limits, not only its strengths.</p>
|
|
| 89 |
<div class="card"><h3>The transcript is evidence, not a convenience</h3><p>Three text files after every conversation: one per language and one with the raw originals — the speakers' words with nothing added by the machine. Timestamps match the audio recording to the second, so any line can be checked at the source or taken home to a professional translator.</p></div>
|
| 90 |
<div class="card"><h3>Never records secretly</h3><p>The transcript and the audio recording are on by default, and the speakerphone says so aloud after Start and again after Stop. You can switch off almost everything — except that announcement.</p></div>
|
| 91 |
<div class="card"><h3>Admits it can be wrong</h3><p>The interpreter introduces itself and tells both sides what to do when a translation sounds odd: say the same thought in different words. It does not hide its limits.</p></div>
|
| 92 |
-
<div class="card"><h3>Lithuanian and Russian translated directly</h3><p>Lithuanian ↔ Russian goes through a direct model, not through English; English and German are direct as well. The interface speaks Lithuanian, English, Russian and German.</p></div>
|
| 93 |
<div class="card"><h3>Everything stays in the room</h3><p>Speech recognition, translation and voices all run on your own PC. No cloud, no accounts, no telemetry, and no internet needed after the first setup.</p></div>
|
| 94 |
<div class="card"><h3>Free & open source</h3><p>GPL v3. Built on open models: Paprika and faster-whisper (the ears), OPUS-MT (translation), Piper (the voices), with our own Lithuanian voice Reginutė from the official Piper catalogue — and a second one on its way.</p></div>
|
| 95 |
</div>
|
|
|
|
| 89 |
<div class="card"><h3>The transcript is evidence, not a convenience</h3><p>Three text files after every conversation: one per language and one with the raw originals — the speakers' words with nothing added by the machine. Timestamps match the audio recording to the second, so any line can be checked at the source or taken home to a professional translator.</p></div>
|
| 90 |
<div class="card"><h3>Never records secretly</h3><p>The transcript and the audio recording are on by default, and the speakerphone says so aloud after Start and again after Stop. You can switch off almost everything — except that announcement.</p></div>
|
| 91 |
<div class="card"><h3>Admits it can be wrong</h3><p>The interpreter introduces itself and tells both sides what to do when a translation sounds odd: say the same thought in different words. It does not hide its limits.</p></div>
|
| 92 |
+
<div class="card"><h3>Lithuanian and Russian translated directly</h3><p>Lithuanian ↔ Russian goes through a direct model, not through English; English and German are direct as well; only German ↔ Russian takes a step through English. The interface speaks Lithuanian, English, Russian and German.</p></div>
|
| 93 |
<div class="card"><h3>Everything stays in the room</h3><p>Speech recognition, translation and voices all run on your own PC. No cloud, no accounts, no telemetry, and no internet needed after the first setup.</p></div>
|
| 94 |
<div class="card"><h3>Free & open source</h3><p>GPL v3. Built on open models: Paprika and faster-whisper (the ears), OPUS-MT (translation), Piper (the voices), with our own Lithuanian voice Reginutė from the official Piper catalogue — and a second one on its way.</p></div>
|
| 95 |
</div>
|