What is vizpilot_ai?

VizPilot AI is a web application that helps users create charts from data files without needing to know which chart type is best for their data. You upload a CSV or Excel file, the system analyzes the structure of your data, and an AI model recommends 3 to 5 suitable chart types along with reasoning, a star rating, and suggested field mappings for axes and color. You then pick a chart, and the app generates an interactive HTML visualization that you can download. The backend is built with Flask and the charts are rendered using Plotly, producing interactive HTML output. The AI recommendation step connects to whichever language model provider you configure. The project supports DeepSeek, MiniMax, and any OpenAI-compatible API, with API keys set either via environment variables or through a settings panel in the UI. Chart types are organized in a registry-driven architecture: all chart metadata lives in a single registry.py file, so adding a new chart type only requires creating its implementation files and registering it there. The rest of the app picks it up automatically. The project ships with over 40 chart types across six categories: comparison charts, time-series charts, distribution charts, geospatial maps, relationship charts, and part-to-whole charts. The project documentation is written in both Chinese and English, and includes screenshots of the home page and the AI recommendation results, along with a FAQ section and a contributing guide for developers who want to add chart types or fix bugs. The project is written in Python (3.7+), backed by Flask, and licensed under Apache 2.0.