SVG infographics from a text stream, AI-friendly
A declarative rendering engine that turns LLM output into polished infographics as it arrives.

What it does
AntV Infographic is a TypeScript rendering engine that takes a custom text syntax and outputs high-quality SVG infographics. It ships with ~200 built-in templates, themes (hand-drawn, gradients, patterns), and a built-in editor for tweaking AI-generated results. You install it via npm, point it at a container, and call render() with a string.
The interesting bit
The syntax is deliberately fault-tolerant so you can feed it a streaming LLM response chunk by chunk and watch the infographic assemble in real time — no waiting for the full JSON payload. It also exposes “skills” (pre-packaged prompts and tools) for Claude Code and Codex, so AI agents can generate, edit, and update infographics without hand-holding.
Key highlights
- Progressive rendering from incomplete syntax — useful for live AI output
- ~200 templates and data-item components out of the box
- SVG-by-default output for crisp scaling and post-editability
- Theme system with presets plus deep customization hooks
- Growing ecosystem: VS Code extension, Obsidian plugin, Slidev addon, React wrapper, Python/Dash bridge, CLI tool
Caveats
- The README calls itself “next-generation” and “declarative” — marketing seasoning you’ll have to filter out
- The AI integration is well-documented for Claude and Codex; other LLM toolchains are roll-your-own
Verdict
Worth a look if you’re building AI-native content tools, slide generators, or documentation systems that need visual polish without a design team. Less compelling if you already have a fixed data pipeline and just need standard charting — this is optimized for text-in, graphic-out workflows.