OpenVids 0.5.1-beta.2

OpenVids 0.5.1-beta.2 — voiceover with your own TTS key.

This is a beta. It is offered only with Settings › General › Updates › "Get beta versions" on, and installs over 0.5.0 and 0.5.1-beta.1. Everything in 0.5.1-beta.1 is included. Voiceover is experimental and may change.

Voiceover (experimental)

• Ask in the chat for a narration: the agent opens a voice setup card, you pick and tune a voice while listening to a sample of your own script, then the agent writes the lines in the model's own syntax, plays the first line in the chat, and generates the rest after "Continue". The lines land on the timeline in a Voiceover audio group, and the music is ducked under them.
• Providers, each with your own key (Settings › Voice, in the editor and on the Projects page): Google Gemini (direct), OpenAI, OpenRouter, ElevenLabs, and your own OpenAI-compatible server (its address is set on the Projects page only). Keys stay on this computer in an owner-only file and are never shown again. OpenVids ships no shared key.
• Generation is paid by you: every paid run asks first with the number of lines, length and estimated cost; identical requests are served from a local cache and never paid twice.
• Gemini: voice catalog with filters, voice design from a description, the vocal tags of Gemini 3.8 (, …), and scene generation (several lines in one request, split by the recognised words). The free tier allows 3 requests per minute for gemini-3.8-flash-tts; OpenVids waits out such limits. On the free tier Google may use what you send to improve its products; in the EEA, Switzerland and the UK its terms require paid services.
• Voiceover tab and clip inspector: edit a line's text, switch between takes without paying again, Regenerate, add lines to the timeline, carve the music under the voice, and add captions from the voiceover (the words of your script, timed by the recording).
• Story Mode: a chapter can carry a narration text; Build Story generates the voice for each chapter and places it at the chapter's start, and narrated chapters get their captions from the voice.

Tested live with Gemini; the OpenAI, OpenRouter and ElevenLabs connectors follow their documentation and have not been run against live accounts yet. Please report problems with Help › Report a Problem.

Turning "Get beta versions" off does not roll the app back: you stay on this beta until a newer stable version is out.