Captions Integrations

Tesseract is an agent-native video editing and motion graphics engine developed by Mirage. It allows AI agents to work directly with professional video concepts including timelines, keyframes, adjustment layers, compositions, typography, timing, effects, footage, and audio. Instead of translating a video project into web technologies such as HTML or React, Tesseract maintains an editable native creative project that agents can manipulate directly. Users can provide footage, reference materials, and a creative brief, then ask an agent to assemble, preview, revise, and render the finished video. The platform supports motion graphics, compositing, animated typography, easing, audio adjustments, and layer-specific revisions within the same project. Tesseract can also let an agent build custom editing controls for elements such as timing, typography, color, framing, and aspect ratio. The runtime operates locally on macOS, Windows, and Linux, with the user’s own computer handling previews and rendering. Tesseract integrates with supported agent environments including ChatGPT and Claude and offers documented interfaces for developers building their own integrations or editors. The engine is free for local use and supports exports up to 4K at 60 FPS, with separate costs applying only to the AI agent or model being used.