Set up your studio
Make the studio yours.
Choose default video settings, review generation providers, and connect your own YouTube channel.
Default video preferences
Generation providers
Failed to fetch
Illustrated motion renderer
AvailableOriginal animated geometry, timed captions, synthesized music and effects. Runs in your browser without external video-model credits.
Built-in AI story & image generation
AvailableAI planning is available immediately. Generated images use the owner’s built-in image credits and service limits. They become animated stills, not model-generated character animation.
Wan 2.2 · fal video provider
Setup neededReal text-to-video character animation, submitted scene by scene. Add FAL_KEY in Functions → Secrets. External provider pricing and limits apply.
AI narration · OpenAI Speech
Setup neededGenerates downloadable speech on the server. Add OPENAI_API_KEY. External speech service availability, pricing and limits apply. Voice recording and uploaded local narration remain free alternatives.
Credentials belong in server Secrets, never in this form or in project storage.
Your YouTube channel
OptionalConnect the built-in YouTube integration, then grant publishing permission with Google’s official OAuth flow. Uploads only start after you review the title, description, visibility, and explicitly approve the video.
Requires GOOGLE_CLIENT_ID, GOOGLE_CLIENT_SECRET, GOOGLE_REDIRECT_URI. App sign-in alone does not grant YouTube access.
What works immediately?
- AI story and connected scene planning
- Animated motion previews & caption editing
- Browser voice preview
- Scene reordering and duration editing
- Account-scoped project saving
- MP4 rendering with available narration, music, and effects
Your work, saved safely
Project metadata and preferences use the built-in per-account key-value store. Uploaded images, audio and video files are public. Do not upload sensitive media.
Signed in as Guest
Setup & honest limits
Use free local models
Wan’s open-source model can be run on your own suitable GPU using its official tooling. Generate scene clips locally and upload them in the editor. Generate voice with local Piper and upload the narration track. This avoids hosted provider fees, but local GPU resources are your responsibility.
Configure model-based generation
Set the server secrets listed under Providers, then refresh configuration. Choose Wan and automatic editing to generate shared scene references (one built-in image credit) and all model clips from your idea; manual editing lets you generate individual clips. Interrupted jobs stay saved and can be resumed with Check generation status. Queue completion is never treated as proof of success. Text-only conditioning cannot guarantee identical characters in every shot.
Export & narration limits
MP4 is recorded natively where supported, otherwise encoded by FFmpeg WebAssembly. Desktop browsers are recommended for 1080p. Browser speech is preview-only; attach real audio for voiced exports. Recorded narration is time-fitted to the final timeline. Scene edits may require new narration.
Configure YouTube OAuth
Enable YouTube Data API in your Google Cloud project, create a Web OAuth client, add your published app origin and exact redirect URI, and set the server secrets. Configure the consent screen and test users. Test the connection on your published app domain, because browser authorization storage is origin-specific. Unverified API projects can be restricted to private uploads by Google. Provider sessions are resolved server-side using random, account-bound references. Google client secrets remain in the editor Secrets panel. OAuth credentials are encrypted at rest using that existing provider credential; changing it requires reconnecting your channel.