FlowCast turns your coding agent into a producer of interactive, narrated walkthroughs — diagrams that draw themselves and code that types itself out, in sync with a voice, scrubbable like video. One self-contained HTML file. No recording, no editing.
Ask in plain language and FlowCast hands back a polished, interactive cast — the kind of thing you'd spend an afternoon recording and editing, generated in one step.
Diagrams draw themselves and code types out in a real editor — crisp at any size, never a blurry screen capture.
Every beat is spoken aloud and captioned, timed to the moment each shape or line appears — so it explains itself while you watch.
Scrub, pause, replay. The whole cast is a single HTML file that opens offline in any browser — share it, embed it, keep it.
Teach a concept as a guided, narrated build instead of a static gist.
Demo a feature or SDK with a walkthrough users can scrub at their own pace.
Onboard contributors — tour the codebase once, hand them a link.
Explain a PR or a system in a cast that travels better than a meeting.
FlowCast is an MCP server. Once it's connected, your AI client gains tools to author and render casts.
Run npx -y @flowcast/mcp setup once — the wizard saves your voice key and wires FlowCast into every AI client it finds.
In any codebase, just say it: “diagram how a request travels” or “walk me through the auth flow.”
Your agent authors the timeline, FlowCast renders it, and hands you a link to a self-contained interactive cast.
No LLM of its own. Your model is the author; FlowCast is the deterministic renderer — your code stays on your machine.
Concept maps and diagrams that draw themselves and connect as the story unfolds.
Files that type out in a live editor with synced narration, precise highlights, and a terminal panel.
Three-dimensional scenes and data stories — like animated charts that build beat by beat.
Runs locally over npx — no install, no account.
Run the wizard, or drop one entry into your MCP config: { "command": "npx", "args": ["-y","@flowcast/mcp@latest"] }.
Casts are narrated, so set a TTS key from OpenAI, ElevenLabs, or Google — your own, stored locally.
“Diagram how data flows through src/.” Open the link it returns and press play.