overview
What ChatSlide does
ChatSlide starts from a source: an uploaded PDF, DOCX, PPTX, TXT, CSV or image, a web page, a YouTube video, a pasted brief or transcript, or a search of research databases such as PubMed, OpenAlex and ClinicalTrials.gov. An agent plans the deck, writes the slides and speaker script, and reviews its own output; you then edit by chat across the whole deck or slide by slide on a canvas, with charts, branding and AI images.
The same deck can become a narrated video, with stock or cloned voices, optional talking-head avatars and a cloned avatar of yourself. Text runs on OpenAI models ChatSlide names (GPT-5.6 Luna and Terra, and GPT-6 Astra on Ultimate); images run on 25+ third-party models such as GPT Image, Imagen 4, Flux, Seedream and Nano Banana; and video can use Veo 3.1 and Seedance 2.5. It is a web app, with 200+ free PDF and PowerPoint tools on the site.
