Pagecast MCP Server

Product walkthroughs in Pagecast begin with an AI-operated browser session. Through record_page, an AI opens a URL and starts capturing video; interact_page then carries out clicks, typing, scrolling, hovering, and navigation in sequence. Calling stop_recording saves the session as WebM. The demo workflow also tracks interaction positions and timing to guide visual effects during export. Developers can rerun a walkthrough after interface changes to update a README demo or product page recording.

Export options separate presentation effects from straightforward conversion. Tooltip mode keeps the full viewport visible while adding magnified insets near interactions; cinematic mode crops and pans between action targets. For recordings without zoom, convert_to_gif uses ffmpeg palette optimization, while convert_to_mp4 produces H.264 video. The record_and_gif tool combines opening a page, recording for a specified duration, and converting the result into a GIF. Saved WebM and GIF files can be found through list_recordings. Destination presets select dimensions and formats for README embeds and social platforms, while plain captures support bug reports and QA work.