Captions Integrations

Tesseract is a creative video suite and local runtime from Mirage built to let AI agents perform professional video editing without relying on conventional editing interfaces. Its project model includes native support for footage layers, keyframes, timing, compositions, adjustment layers, graphics, effects, typography, and sound. This allows agents to manipulate the underlying components of a video directly rather than recreating editing behavior through webpage or React-based representations. Users can start with existing footage, images, audio, screen recordings, or generated media and provide an agent with a natural-language creative brief. The agent can then cut footage, create motion graphics, composite elements, adjust pacing, animate text, shape audio, and render a preview or final export. Because Tesseract keeps projects editable, users can request a change to a specific title, clip, keyframe, effect, or audio element while preserving the rest of the project. The platform also allows agents to create custom manual editing interfaces with controls for timing, typography, color, framing, and other project parameters. Tesseract runs locally on macOS, Windows, and Linux, while supported hosted environments can also operate the engine where available. It can be used with ChatGPT, Claude, Muse, Grok Bot, Instinct, and other compatible agents through plugins, skills, or documented interfaces. Tesseract supports local preview and rendering, including 4K output at up to 60 FPS, and does not require an Adobe subscription or Mirage account for local use. Mirage also uses Tesseract as the underlying engine for Captions, while Captions remains a separate application and cloud workspace with additional creation services.