Audio

Say their name in your voice.

SMS and text-to-speech scripts accept {{first_name}}, {{last_name}}, {{company}}, and any custom field stored on the contact. Bulk sends render each contact on their own. Uploaded audio still works and stays one credit per delivered voicemail.

Voice cloning turns on when ELEVENLABS_API_KEY is set. Until then the workspace says “Voice cloning: connect a voice provider.” A clone needs a WAV or MP3 sample of at least 30 seconds and a checkbox that you own the voice. That consent text and the time are stored with the voice. Personalized speech is cached and rendered through a concurrency queue so a large send does not stampede the provider. Personalized voicemail defaults to 2 credits. Set DROPAGENT_AI_VOICE_CREDITS and DROPAGENT_STATIC_VOICE_CREDITS to change the amounts.ELEVENLABS_CONCURRENCY defaults to 2.

Test mode does not call the voice provider and does not spend credits. Pass voiceCloneId and script on POST /api/v1/drops.