flint-chart

Why we included this project

Anyone who has asked an AI agent to produce a chart knows the usual outcome: technically correct and visually forgettable. Flint attacks that problem by treating most chart design decisions as the compiler's job, not the prompt's. You write a compact spec saying what each field means and roughly what you want, and the library handles the scales, spacing, labels, and layout before emitting working code for Vega-Lite, ECharts, Chart.js, Plotly, or native Excel charts. That appeals both to teams wiring chart generation into an agent's output and to developers who would rather describe a chart than tune a charting library by hand. The bundled MCP server extends the same spec workflow into a chat or coding environment, letting a model create, validate, and render charts directly. For shops with a strong visual identity, reusable themes such as Economist or Swiss keep brand styling consistent across every generated chart.