Skip to main content

​
What you will build

  • A single page where you type text, pick a voice, and generate a clip.
  • Audio generated by Deepgram’s Aura text-to-speech API.
  • Inline playback plus a download option.

​
What you need

A Deepgram account and API key saved as DEEPGRAM_API_KEY in your project environment (Project Settings → Environment or .env.local).
  • Sign up or log in at deepgram.com → Console (console.deepgram.com).
  • Create a project if prompted, then go to API Keys → Create API Key.
  • Choose a name (e.g., mistle-voiceover), keep scopes broad (default works) or ensure Text-to-Speech (Speak) access, and create the key.
  • Copy the key once (you won’t see it again) and store it in your env as DEEPGRAM_API_KEY.

​
Create Project Prompt

Create a voiceover app that turns typed scripts into audio clips with Deepgram.

​
Plan A Feature Prompt

Build a voiceover app on a single page. The user will write a text and send it into the input form and will receive a voiceover audio back with the Deepgram API.

​
Ideas after the initial build

  • Ask Mistle to scan the Deepgram TTS docs (https://developers.deepgram.com/docs/tts-rest) and propose UX/feature tweaks tailored to your stack.
  • Add voices selection
  • Offer language options
  • Keep a recent-history list with play/download links so users can compare takes without retyping.