AI
The AI workspace provides paid generation and editing tools inside Beutl. Available models, input types, and options depend on the selected task and model.
Tab characteristics
- Open by default: No
- Allow multiple instances: Yes (each tab keeps its own work)
How to open
Open View → Tools → AI, or choose a task from the AI menu while a scene is open.
Sign in to your Beutl account. If your account cannot use AI, the workspace offers Pro-plan guidance. The plan button opens account management in your browser; returning to Beutl refreshes the account information.
Sections
| Section | Purpose |
|---|---|
| Image generation | Generate an image from a prompt |
| Image editing | Remove a background or object, upscale, restyle from a reference image, or outpaint |
| Video generation | Generate video with references supported by the model |
| Video editing | Edit or extend a video, or apply motion using a video and a person/character image |
| Subtitles | Generate subtitles using a selected subtitle template |
| Jobs | Follow AI requests and their results |
Video generation can use the current scene frame or first/last-frame inputs where supported. General references may include images, videos, or audio; supported combinations are shown for the selected model. Editing an existing video uses its duration rather than a separate duration choice.
You can drag local files onto the corresponding input area: reference images, a source image or video, first/last frames, a character image, supported video-generation references, or subtitle imports. The active task and model determine which file types and how many references are accepted.
Subtitles take a Text template: the default look, or a saved template that holds text, whether of a TextBlock, an element, or several elements. Its formatting, effects, animation, and placement apply to every subtitle; see Subtitle files for how the elements of a template fit each cue. Import captions reads SRT, WebVTT, and ASS/SSA files, including UTF-16 files with a byte order mark and files in the system's legacy code page.
Image editing results offer Add to scene and Save to file actions. Review the result before adding it to your project.
Related documents
- Node Graph — Connect AI generation steps in a graph
- Editing with AI Agents — Connect an external coding agent through MCP