Zing quickstart

What you need

You need Python 3.11 or newer, uv, and FFmpeg with libx264 and libass on your PATH. Clone the repo, run uv sync --all-extras, then uv run zing studio. For a smaller install, pick only the extras you want: study (transcription and OCR), mcp, render, memes, fonts, layout (faces), color and audio. The first time you edit by speech, Zing needs a speech model download. Studio and doctor name it and ask before anything downloads.

Get the checkout ready

git clone https://github.com/ryanbiddy/zing
cd zing
uv sync --all-extras
uv run --frozen zing doctor
uv run zing studio

Run these commands in your terminal. Doctor reports missing tools and feature dependencies. Review download requirements before starting a study.

Connect your AI client

uv run --frozen zing serve-mcp --print-config

This prints settings with your environment's interpreter path. Merge the applicable settings into your MCP client, restart its connection, then ask it to call zing_status.

Check the connection

Ask your connected client to call zing_status with no arguments. It reports tool availability, workspace stats, running or failed study jobs, and available prompts.

{
  "name": "zing_status",
  "arguments": {}
}

What touches the network

Studying, rendering and memes run locally by default. Reference URL downloads require explicit consent. Requested model and font downloads are pinned by URL and SHA-256, cached locally and reported. Uoink is reached only through a peer you configure on your machine. Zing has no telemetry or analytics.