Recording and Transcription

Choose transcription and output languages

Distinguish the language spoken in a recording from the language used for AI replies.

Spoken language

In Settings > Account, use Transcription Language for the language spoken in the recording. Leave detection on Auto when appropriate, or select the known language if detection is incorrect.

Do not select a different language as a substitute for an unavailable one. Check the current selector and language coverage for the relevant processing path.

AI output language

Output Language controls the language used for AI writing, such as replies and generated notes. It does not change what language was spoken in the recording.

For a translated transcript, use the recording’s translation action and choose the target language. For translated speech, see dubbing.

Coverage and accuracy

The documented provider union covers 100+ languages. This is a deduplicated coverage figure, not a guarantee that every live-caption, transcription and translation path supports each language equally.

The maintained language list and provider evidence are on Accuracy and Methodology. Test important content with a representative sample and review the result. The Help Center does not maintain a separate copy of that list.