ittybitty MCP server
Enables AI agents to generate narrated videos from topics or scripts, with stock footage, home videos, or local AI clips.
README
ittybitty
Topic or script in → narrated video out. Shorts for TikTok and Reels, or longer chaptered explainers — with stock footage, your Jellyfin/Plex home videos, or local AI clips.
Alpha — under construction. APIs, pipelines (R3 screening, R9 talker), and the Windows installer may change between releases. Not production-ready. See alpha release checklist before going public or tagging.
Three ways to use it (pick one):
| Windows app | Web dashboard | MCP server | |
|---|---|---|---|
| For | Everyday use on your PC | Browser UI while you develop or self-host | Cursor, Claude, other MCP clients |
| You get | Installer + desktop shortcut; backend starts automatically | Generate, Jobs, Depot, Settings, Help | videogen_* tools for agents |
| Start | Download installer | Run locally or use the app above | Connect MCP |
Same engine under all three: Python backend on port 11054, React dashboard on 11055 when running the dev stack.
Contents
| Windows app | Installer — no Python required |
| Web dashboard | Browser UI (dev or bundled in desktop app) |
| MCP server | Agent automation |
| What you need | Keys and optional extras |
| Footage sources | Where B-roll comes from |
| vs MoneyPrinterTurbo | Why pick ittybitty |
| More help | Config, troubleshooting, API docs |
| Marketing site | GitHub Pages one-pager (also website/ locally) |
Windows app
Best if you just want to make videos.
- Download the latest pre-release installer from Releases (tag
v0.2.0-alpha.*or newer) - Run the installer → launch ittybitty from Start or your desktop shortcut
- In Settings, add a free Pexels key → Generate with a topic or paste a script
The installer bundles the web dashboard and Python backend — no separate Python or Node install.
Install folder: %LOCALAPPDATA%\ittybitty\
If the backend fails to start: %LOCALAPPDATA%\ai.fleet.ittybitty\logs\backend-spawn.log
Web dashboard
The React UI: Generate, Plan (mid-length), Jobs, Depot, Publish, Settings, Help, Logs.
Already using the Windows app? You're in the dashboard — nothing else to install.
Running on your dev machine (live reload, for contributors):
git clone https://github.com/sandraschi/ittybittyvideos.git videogen-mcp
cd videogen-mcp
.\start.bat
Open http://127.0.0.1:11055. The script starts the dashboard and the API together.
Needs Python 3.10+, FFmpeg, and Node.js — see INSTALL.md.
MCP server
For AI agents that should generate or plan videos over the Model Context Protocol.
- Start the backend (Windows app running, or
.\start.bat -BackendOnly, oruv run python -m videogen_mcp.server) - Point your MCP client at
http://127.0.0.1:11054/mcp
Example prompts: "Generate a 45-second video about Vienna coffee culture" · "Plan a 5-minute sourdough tutorial" · "Call videogen_help first"
Tool list and REST API: docs/TOOLS.md (16 MCP tools; catalog at /api/v1/tools) · Claude Desktop / MCPB: INSTALL.md
What you need
| For | You need |
|---|---|
| Most workflows | FFmpeg on PATH + free Pexels API key |
| Topic → script | DeepSeek or OpenAI key, or paste your own script |
| Home videos as B-roll | Jellyfin or Plex URL + token (Settings) |
| Local AI clips (GPU) | CUDA ~24 GB + .\start-localgen.bat |
The in-app Help page walks through each step.
Footage sources
Pick one in Settings → Footage:
- Pexels — free stock (default, no GPU)
- Jellyfin / Plex — cut clips from your library (vacation, pets, …)
- Veo / Omni — Google cloud (config)
- LocalGen — Wan 2.2 on your GPU
Finished videos land in ./output/ and appear in Depot.
How this differs from MoneyPrinterTurbo
MoneyPrinterTurbo popularized topic → short video. ittybitty targets the same job with fleet tooling: agents, longer formats, your own libraries, and a path to local GPU — still alpha, so expect rough edges.
| MoneyPrinterTurbo | ittybitty (alpha) | |
|---|---|---|
| How you run it | Web UI | Windows app + web dashboard + HTTP MCP |
| Length | Mostly shorts (~60s) | Shorts + 3–15 min chapters (planner + videographer rules) |
| Footage | Stock APIs | Pexels + Jellyfin/Plex + LocalGen (Wan 2.2) + Veo/Omni |
| Agents | — | videogen_* tools for Cursor, Claude, fleet MCP clients |
| Library & jobs | — | SQLite depot, job history, publish handoff |
| Edit intelligence | Assembly / concat | Hook, pacing, B-roll, transitions; R1 karaoke subs; R2 beat snap + ducking; R3 screening (experimental) |
| Architecture | Monolith | Plugin registry — swap LLM, stock, and TTS providers |
| Desktop | — | NSIS installer (Windows, backend bundled) |
| Status | Mature OSS | Alpha — checklist |
中文说明与中国本地栈(通义千问、CosyVoice 等):README-zh.md
Sample output
Demo reel coming soon (GSD puppy short). Quick test render:
py scripts/smoke_render.py
More help
| Doc | When to read it |
|---|---|
| INSTALL.md | All install paths, MCPB, verification |
| docs/CONFIGURATION.md | Env vars and providers |
| docs/TROUBLESHOOTING.md | Something broke |
| docs/ALPHA-RELEASE-CHECKLIST.md | Going public / alpha tag |
| docs/PROMPT-DIRECTOR.md | R10 trope templates (mermaid) |
| docs/EXEMPLARS-RESEARCH.md | Viral Short formats + mid expansion |
| docs/TOOLS.md | MCP tools and REST API |
| docs/DEVELOPMENT.md | Tests, Tauri build, CI |
| SPEC.md | Architecture and roadmap |
| CHANGELOG.md | Version history |
Fleet docs: mcp-central-docs/projects/ittybitty
MIT · sandraschi · v0.2.0
Repo folder videogen-mcp, Python package videogen_mcp — kept for MCP tool compatibility.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.