Install
keyline is a free, source-available MCP server that lets AI agents (Claude Code, Cursor, Codex, VS Code) design social posts, display ads, banners, flyers and video at every size from one JSON scene. It measures text fit and overflow before rendering, and renders with Skia instead of a headless browser.
Claude Code (macOS or Linux): the plugin downloads keyline and checks it against the release's SHA256SUMS, or uses keyline-mcp from your PATH.
/plugin marketplace add keyline-dev/keyline
/plugin install keyline@keyline
Claude Desktop (Mac with Apple silicon, or Windows): download keyline-mcp-<version>.mcpb from the latest release and double-click it. Designs and their renders go in ~/keyline, a folder per design, where you can also put the images and logos Claude should use. To work in folders of your own instead, open Settings → Extensions → keyline → Configure, add them, and click Save.
Other clients: install, then add
| Install | |
|---|---|
| macOS (Apple silicon) | brew install keyline-dev/tap/keyline-mcp |
| Linux (amd64, arm64) | sudo apt install ./keyline-mcp_<version>-1_amd64.deb (or _arm64), from the latest release; tarballs are there too |
| Windows (x64) | Unpack keyline-mcp-v<version>-windows-amd64.zip and put its folder on your PATH |
| Docker | Use docker run -i --rm -v keyline:/data ghcr.io/keyline-dev/keyline-mcp as the client's command; the :stills tag leaves out ffmpeg |
| From source | cargo build --release (Rust stable; on Linux also libfontconfig1-dev libfreetype6-dev) |
Every release has a SHA256SUMS file and build attestations: gh attestation verify <file> --repo keyline-dev/keyline. Then add keyline to your client. The command is keyline-mcp; if the client can't find it, give its full path (which keyline-mcp).
Codex is OpenAI's coding agent, included in ChatGPT plans. Its app, CLI and IDE extension share one setting:
codex mcp add keyline -- keyline-mcp
ChatGPT itself, like grok.com and the xAI API, connects only to remote servers, so it can't start keyline, which runs on your machine.
Any client that starts local (stdio) MCP servers works, Cursor, VS Code and Gemini CLI included. Add a server named keyline whose command is keyline-mcp, in the client's MCP settings or with its mcp add command. Most clients take this JSON:
{ "mcpServers": { "keyline": { "command": "keyline-mcp" } } }
Claude Code without the plugin (and on Windows): claude mcp add --scope user keyline -- keyline-mcp. Claude Desktop without the .mcpb: Settings → Developer → Edit Config, with the full path (/opt/homebrew/bin/keyline-mcp), since it doesn't read your shell's PATH.
Then
Ask your agent for a design: "Make a vote-by-mail flyer with this photo, in 1080×1350, 1200×1000 and a 300×600 skyscraper."
- Your files: each design gets a folder,
~/keyline/<name>/, holding its scene andrenders/; the agent adds images, video and templates in it by path (cheaper than sending their bytes) and picks up a design by name in a later session. The Claude Code plugin uses the project folder instead. Elsewhere,--folder ~/work ~/brandreplaces~/keyline: designs go in the first folder, and files are read in all of them. Flags go after the command (claude mcp add keyline -- keyline-mcp --folder ~/work), or inargsin a JSON config. In Docker, designs are in the/datavolume; mount other folders and pass them. - Video needs ffmpeg on the PATH; everything else, animated PNG and GIF included, works without it.
- Options: every setting is a flag (
--folder,--no-motion,--data,--fonts,--renderer,--ffmpeg,--encoder); seekeyline-mcp --helpand Tools and replies. - Without an agent:
keyline-mcp render scene.json --out renders/renders every size and exits 1 on a!defect; in GitHub Actions,uses: keyline-dev/keyline@v0does it for a repo's scenes (details). - GPU on a Linux server needs Vulkan drivers (NVIDIA's, or Mesa); without a GPU, keyline renders on the CPU.
Also as Markdown.