EduAIttractive

We produce short educational videos from a wide range of sources. We save time for healthcare experts (and others).

Users can upload a public video URL (e.g. a YouTube podcast), their own audio and video files, documents (PDF, PPTX, HTML), images and charts, webpages or their own video description to the tool. The sources are converted into text form and processed by LLM models to create a video script. The user sees an overview of the individual scenes in the video and can edit the text within them in real time. The text is sourced back to the input materials (document, page, time, citation) or, where applicable, marked as having been generated by the model.

The video is rendered deterministically by Remotion (React), making it reproducible and allowing each scene to be edited manually. A source is stored for every statement in the video.

You can choose between various predefined visual styles, or customise them or create new ones. You can perform basic edits on the visual elements contained within the video directly – such as duplicating, deleting and moving them. In ‘Creative’ mode, the AI programmes the animation for each scene individually. Changes are versioned and can be undone. A work-in-progress video can be saved as a draft and returned to later, or the work can be deleted or published. This action adds the video to a shared gallery, where it is available to other users of the app, or it can be downloaded as an MP4 file. Most of the work on the video can be carried out either through manual adjustments or by giving a prompt to the AI.