Microsoft Research released Flint, an open-source visualization language designed to help AI agents generate expressive charts from compact, human-editable specifications.

The language addresses a core problem in AI-driven visualization: existing tools either produce uninspiring results from simple specs or require verbose, error-prone configurations for polished output. Flint offers a middle path by using semantic data types to guide design decisions automatically.

How semantic types drive chart quality

Flint leverages semantic data types to express the meaning of data fields — whether a column represents dates, prices, percentages, or rankings. The compiler uses this information to choose appropriate scales, baselines, formatting, and color schemes without explicit user configuration.

The system automatically manages sizing, spacing, labels, and layout so charts remain readable as data complexity changes. A single Flint specification can compile to multiple backends including Vega-Lite, Apache ECharts, and Chart.js without rewriting.

Microsoft's research team, led by Senior Researcher Chenglong Wang, compared Flint against DirectVL, a baseline approach that generates full Vega-Lite specifications directly. The study found Flint better suited for LLM-based chart generation because semantic types are easier for models to infer than low-level visualization parameters.

"Field names, value patterns, and common data knowledge can help an agent recognize whether a column represents a date, price, percentage, country, ranking, or correlation," the researchers wrote.

The open-source project includes the flint-chart library and flint-chart-mcp server, enabling agents to create, validate, and render charts directly in chat or coding environments. This integration allows AI systems to produce publication-ready visualizations without managing complex backend-specific parameters.

Flint represents Microsoft's broader push into AI tooling, joining efforts across the company to make AI development more accessible. The language is available now on GitHub with documentation and examples for developers building AI-powered data visualization systems.