AI Speaker usage guidelines

This page covers how to use AI Speakers in Descript, including text-to-speech limits, regeneration allowances, and voice clip guidelines for each subscription plan. Find out what’s included with your plan and how to maximize Descript's AI-powered voices.

AI Speaker consent statements must be recorded in English. To learn more about usage limits for AI features, visit our Understanding AI Limitations, Features, and Usage Tracking page.

  • AI Speakers Availability by Plan: You can use AI Speakers with any Descript subscription, but the usage limits vary by plan:
    • Free Plan: 5 minutes of text-to-speech, 5 regenerations per month.
    • Hobbyist Plan: 30 minutes of text-to-speech, 10 regenerations per month.
    • Creator Plan: 120 minutes of text-to-speech, 100 regenerations per month.
    • Business Plan: 300 minutes of text-to-speech, unlimited regenerations.
    • Enterprise Plan: Unlimited text-to-speech and regenerations.
      • Access to the Overdub API, which was previously limited to enterprise plans, is no longer available.

  • Project Location and AI Speaker Access: If you have multiple drives, make sure to create or move your project to the paid drive to access paid-level AI Speaker limits.
  • Voice Clip Limits: A voice clip is defined as:
    • Any text segment (250 characters or fewer) with a pause of 2 seconds or more between typing.
    • A sentence (250 characters or fewer) ending with sentence-ending punctuation.
    • A string of 250 characters or fewer without sentence-ending punctuation.
  • Text-to-Speech (TTS) Paragraph Length: For best results, keep paragraphs under 1800 characters. Shorter paragraphs will generate speech faster, so breaking up longer text is recommended for quicker processing.
  • Text-to-speech (TTS) Calculation: The length of generated AI speech counts towards TTS minutes. TTS usage works within a "bucket" system that doesn’t reset when you upgrade. For example, if you're on a plan with 30 minutes per month and upgrade to a plan with 2 hours per month, the minutes you’ve already used will count towards the new plan’s limit. Users may occasionally exceed their TTS limit slightly before Descript restricts further usage. When this happens, Descript will restrict usage until the next monthly reset.
    • Note: Every time the AI speech is regenerated or the AI speaker is changed, it will count toward your TTS minutes. For example, editing even a section in Write mode will result in additional TTS usage as the speech is regenerated.
  • Supported Text-to-Speech (TTS) Languages: You can generate TTS using a custom AI Speaker in the following 19 languages: Catalan, Finnish, Lithuanian, Slovak, Croatian, French (FR), Malay, Slovenian, Czech, German, Norwegian, Spanish (US), Danish, Hungarian, Polish, Swedish, Dutch, Italian, Brazilian Portuguese, Turkish, English (US), Latvian, and Romanian.

After reaching the voice clip limit for your plan, you will receive an in-app notification and will be blocked from generating more AI Speech.