ElevenLabs Dialogue V3: AI Dialogue Audio Generator

Use it to hear a short scripted exchange, such as a podcast introduction, interview rehearsal, or character scene. Assign a voice to each dialogue line, then listen to the result to check whether the delivery suits your script.

You can add dialogue rows until the form reaches 5 rows. Each row needs text and a selected preset voice. Stability and an optional language code apply to the request. Choose the voice in each row; the separate Voice preset selection does not set a row's voice.

Keep the combined script within 5000 characters. Rivya counts each voice name, the ': ' separator, the dialogue text, and the line breaks between rows. This is a script limit; the credit estimate counts only each line's text after trimming its ends.

When you submit a request, Rivya reserves 14 credits per 1000 characters across the dialogue text, rounded up to a whole credit with a minimum of 1. Each line's leading and trailing whitespace is excluded from this estimate. On successful completion, a valid usage report that is no higher than the reserve determines the final charge, rounded up to a whole credit; any difference is refunded. The estimate is not a fixed final cost.

When all lines are for the same narrator, a single-voice text-to-speech model can be a simpler setup. Use Dialogue V3 when you want to assign a voice to each turn in a scripted exchange.

Try a short exchange between speakers before submitting a longer script. Select a voice for each line and listen to the resulting audio to decide whether those voices suit the scene.

Write the spoken text in the dialogue rows and select a preset voice for each row. The separate prompt field does not supply the spoken dialogue. This workflow takes a script and produces audio; it does not use uploaded voice samples or video files.

Check that every row has text and the voice you want for that line. Review Stability and leave Language code empty if you do not need to send one. Listen to a result before deciding whether the settings suit your script.

You can read this page without signing in. Sign in to submit a dialogue request. Rivya pre-debits 14 credits per 1000 characters of trimmed dialogue text, rounded up with a minimum of 1. A successful task can settle at a lower charge and return the difference; check its credit status for the final amount.

With a Rivya API key, send params.dialogue as an array of objects with voice and text. Use a preset voice name for each object. Supply the dialogue directly: default_voice does not fill missing line voices, and prompt does not supply the dialogue. Set stability and language_code inside params if needed.