Set up bilingual captions for a call.

Select and check your audio source

Choose the device or application you want to hear in the main audio selector. Keep your meeting’s playback routed to that source. If captions remain empty, check the selected source, meeting output and whether speech is actually playing. Avoid capturing unrelated applications or a private conversation during troubleshooting.

Choose the spoken and target languages

Open Spoken languages and select a combination supported by the installed recognizer and translator. English–Spanish is the included starting combination. Other supported pairs may require downloaded packs; model-language availability checks decide which choices are usable. An interface language is not a promise of speech recognition support.

Prepare a directional glossary

Open Glossary, choose the source and target language, then add the term and the wording you want in the target. Check both directions when you work bilingually: a preferred English-to-Spanish term is not automatically an approved Spanish-to-English translation. Glossaries guide terminology; they cannot guarantee recognition or repair speech that was never recognized.

Read, review and export

Compare original and translated captions, especially names, amounts, dates and negation. Scrolling away pauses following; return to the bottom or use the return-to-live arrow to resume. Use the available copy or save controls only when you want an export. Transcript autosave is optional; handle any saved conversation according to the participants’ permissions.

Report a useful problem

Record the app version, selected models, language direction and the step that failed. Provide a short invented example that reproduces the problem where possible. Do not send payment details, customer recordings or private transcripts by default. More complete first-run and updater workflows are still under development.