voiceme-ai

Inspiration

Creators often lose hours re-recording voiceovers because of small mistakes, background noise, or changing recording conditions. We wanted to make narration fast without losing the creator's own voice and identity.

What it does

voiceme-ai lets you record or upload a short voice sample, explicitly consent to voice cloning, write a script, and generate AI narration in your own voice. You can adjust speed and pitch, preview the result, and download the audio.

The public demo runs in mock mode, so anyone can experience the full product flow without uploading personal audio or needing an API key.

How we built it

We used Codex and GPT-5.6 to research the voice-cloning workflow, shape the consent-first product experience, implement the responsive voice studio, add a MiniMax Voice Clone/TTS integration path, and create a reliable mock demo mode.

The app uses vanilla HTML, CSS, and JavaScript for the studio UI, Node.js and Express for the API layer, Web Audio APIs for recording, and Vinext-compatible deployment for the public demo.

Challenges we ran into

The main challenge was balancing a convincing demo with privacy and reliability. Voice cloning requires careful consent handling, API credentials must stay server-side, and browser-recorded audio needs to be converted into a provider-friendly format.

We also handled large provider file IDs safely and added a mock generation path so judges can try the product even when no external TTS credentials are configured.

Accomplishments that we're proud of

• Built a polished Korean-first voice-cloning studio.
• Added explicit consent, preview, download, and delete controls.
• Made live and mock modes share the same user experience.
• Created a public deploy that works without API keys or personal audio uploads.
• Kept the project open source and easy to run locally.

What we learned

Voice AI is not only about synthesis quality. Trust, consent, privacy, and a clear creator workflow are equally important. We also learned that a strong fallback path makes an API-backed project much easier to evaluate and demo.

What's next for voiceme-ai

Next, we plan to add opt-in local inference, streaming generation, reusable voice profiles, subtitle timestamps, and creator presets optimized for short-form video narration.

Built With

• codex
• express.js
• gpt-5.6
• javascript
• minimax
• node.js
• vinext
• web-audio-api