Chat chips for audio and video: what each one does
Attach audio and four chips appear: Recognize speech, Translate audio, Content analysis, Clone voice. For video there is one: Translate video.
Updated November 2, 20263 min read
Chips are buttons above the message field that process an attached recording in one tap, without writing a request. They appear as soon as you attach an audio or video file with +. For photos and documents there are no chips: write what to do in the message.
What each chip does
| Chip | File | What you get |
|---|---|---|
Recognize speech | audio | The text of the recording as a chat message |
Translate audio | audio | The app asks for the target language, then shows the translation and the original text. The answer is text only, without voice-over |
Content analysis | audio | Four sections: Topic, Key points, Tone, Summary |
Clone voice | audio | The voice from the recording is saved to Settings → My Voices under the file name. The chat answers “Voice '…' cloned and available in Text-to-Speech” |
Translate video | video | The app asks for the target language and sends the video to Video Translation. The chat answers “Video sent for translation. You will be notified when ready.” The finished video appears in the Video Translation history |
If several files are attached, the chip takes the first one and removes the rest from the message.
How to use a chip
In the Chat tab, tap + and choose an audio or video file.
Tap a chip above the message field. You don't need to type anything.
For Translate audio and Translate video, choose the language in the window that opens.
Wait for the answer. After that you can continue the conversation: ask for a shorter version, a translation into another language or a list of words from the text.
When to open the tool instead of the chip
The chips give a quick result in the chat. The full tools on the Tools tab do more:
To hear the translation in the speaker's own voice, use Audio Translation. The chip only translates the text.
For speaker labels, timestamps and a summary of a long recording, use Text Transcription.
To record a voice sample while reading a prepared text and to choose the cloning engine, use My Voices.
To choose the dubbing engine and lip sync, open Video Translation directly.
How much the chips cost
Each chip costs about the same as the tool behind it: speech recognition and text translation cost fractions of a token per minute of audio, cloning a voice costs about 1 token, and video translation is charged at the Video Translation rate. Exact prices are in token prices for each tool.
Suggestions in an empty chat
A new chat without files shows other buttons: suggestions such as Translate a voice note, Dub a video in my voice, Translate a photo or Download a video. A suggestion sends a ready request, and the AI explains what to send next. The set of suggestions changes and depends on what you have used before.
Frequently asked questions
I attached a video but there is no Recognize speech chip. Why?
For video there is only Translate video. To get the text of a video, write “Transcribe this video” in the message, or use Video to Text.
The chat said “No speech detected in the audio”. What now?
The recording has no recognizable speech: it is silent, too quiet or only music. Check that the right file is attached and that the speech is audible.
Contact support
Describe the problem and we will reply by email. Add the file name or job ID if you have one.