About ChalkTalk
ChalkTalk is an MCP server for creating narrated tutorial videos from a conversation, aimed at developers working in editors and AI tools. It structures videos as scenes, validates drafts, and renders preview frames for review before export.
Key features
- 01Create and edit video projects
- 02Search projects and templates
- 03Validate structure and pacing
- 04Render preview frames
- 05Generate approved voiceovers
Use cases
- 01Draft a narrated code walkthrough
- 02Explain a pull request in video form
- 03Review a tutorial's pacing
- 04Preview scenes before exporting
MCP support
Provides an MCP endpoint for AI clients to create, explore, validate, and preview video projects. Users sign in in a browser on the first tool call.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://chalktalk.dev/api/mcpclaude mcp add --transport http chalktalkhq-chalktalk https://chalktalk.dev/api/mcp{"mcpServers":{"chalktalkhq-chalktalk":{"type":"http","url":"https://chalktalk.dev/api/mcp"}}}Open in Cursor{"servers":{"chalktalkhq-chalktalk":{"type":"http","url":"https://chalktalk.dev/api/mcp"}}}Open in VS Code{"mcpServers":{"chalktalkhq-chalktalk":{"type":"http","url":"https://chalktalk.dev/api/mcp"}}}{"mcpServers":{"chalktalkhq-chalktalk":{"serverUrl":"https://chalktalk.dev/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEChalkTalk
Turn a conversation into a narrated tutorial video - without leaving your editor.
ChalkTalk writes the video as a structured document: scenes for code, diagrams, charts, CLI sessions, and maths, each with its own narration. Ask for a walkthrough of the function you just wrote and it drafts the scenes, checks them, and renders preview frames it can actually look at. You review and export in ChalkTalk.
Install
| Client | How |
|---|---|
| Claude Code | Plugin - two commands below |
| Codex | Two CLI commands below |
| Cursor, VS Code | One click from the install page |
| Claude Desktop, claude.ai, mobile | Settings → Connectors → Add custom connector → server URL |
| ChatGPT | Settings → Security and login → Developer mode, then Plugins → + → Public endpoint → server URL |
| Windsurf, Antigravity | Config snippet below |
| Anything else | The install page has a prompt you can paste into your own agent |
Claude Code
/plugin marketplace add chalktalkhq/plugins
/plugin install chalktalk@chalktalk-plugins
Then run /mcp to sign in.
Codex
codex mcp add chalktalk --url https://chalktalk.dev/api/mcp
codex mcp login chalktalk
Windsurf, Antigravity - merge into your MCP config. Windsurf keeps it at
~/.codeium/windsurf/mcp_config.json; Antigravity exposes it under
MCP Servers → Manage MCP Servers → View raw config.
{
"mcpServers": {
"chalktalk": {
"serverUrl": "https://chalktalk.dev/api/mcp"
}
}
}
Server URL, for anything that just asks for one:
https://chalktalk.dev/api/mcp
No API key. You sign in once in the browser on the first tool call.
Try it
- "Make a 90-second video explaining this pull request."
- "Turn this function into a code walkthrough with narration."
- "Start from the binary search template and show me the first scene."
- "Check my project for pacing problems and show me frame 3."
What it can do
| Author | Create, update, duplicate, and delete video projects |
| Explore | Search and fetch your projects and the template catalogue |
| Check | Validate structure, report pacing and writing problems |
| See | Render preview frames the agent can inspect before you commit |
| Narrate | Generate voiceover from narration you have approved |
Rendering runs in ChalkTalk rather than over the connector, so you review the video before spending render time on it.
Links
This repo is generated from the ChalkTalk app repo - open an issue rather than a pull request against these files.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.