Articles · Productivity
Turn a script into a video without a paid voice API
A talking explainer should not require a studio, a subscription voice, or a cloud that keeps the script. The Script to Video Maker splits the draft into scenes, lets you attach images, adds captions and motion, then exports a WebM file. Auto Voice is optional: when you want spoken English, Kokoro runs in the browser. When you already recorded Urdu, Arabic, Hindi, or any other language, you upload that file. When the pictures and captions are enough, you choose No Voice.
Why this is worth doing in the browser
Most “AI video” products send the script to a paid speech vendor. That is the wrong default for a draft, a lesson, or a private brief. DevNestro keeps Auto Voice on the device. The first visit may download a large Kokoro model from Hugging Face into the browser cache. After that, generation does not post your sentences to a speech API. Urdu and other scripts still work for captions and timing. Auto Voice simply stays off until a matching local model exists; it will not fake Urdu with an English speaker.
How to use the tool
Write the script with a blank line between scenes. Add images if you have them. Pick ratio, transition, and motion. On the Narration step choose Auto Voice, Upload Voice-over, or No Voice. Preview a short English sample before you generate every scene. Sync scene lengths to the audio, mix optional music with ducking, then download. Share for 24 Hours uses the existing File Share path only if you confirm an upload.
- Keep Auto Voice for English scripts; upload or skip voice for Urdu and other unsupported languages.
- Prepare the voice engine only when you intend to use Auto Voice so the model is not downloaded on every visit.
- Generate scene by scene, retry a failed scene, and regenerate only the scenes you edited.
- Preview pause and seek before export. The downloaded file is the mix of pictures, captions, narration, and music.
Privacy
Auto Voice runs in your browser. The script and generated speech are not sent to a speech service. The voice model may download from a trusted host on first use. Optional sharing uploads only the finished video you choose to send through File Share, which expires after 24 hours.
If the job is a script plus pictures, do not rent a cloud voice for the first cut. Open Script to Video Maker, choose Auto Voice or no voice, and download the file from this device.