Describe your data. Get production charts. Your data stays local.
Quickstart •
API Reference •
Examples •
Changelog
NVEIL is an AI-powered data visualization toolkit. Write one line of natural language, and NVEIL processes your data and generates publication-ready visualizations — no chart code, no hallucinations, no data leaving your machine.
import nveil
nveil.configure(api_key="nveil_...")
spec = nveil.generate_spec("Revenue by region, colored by quarter", "sales.csv")
fig = spec.render("sales.csv")
nveil.show(fig)
From your shell
After pip install nveil the nveil command is on your $PATH:
export NVEIL_API_KEY=nveil_...
nveil describe sales.csv
nveil generate "Revenue by region, colored by quarter" \
--data sales.csv --format all --explain
nveil render chart.nveil --data new_sales.csv
For AI agents (Claude Code / Claude Desktop / Cursor / Codex / …)
NVEIL ships first-class integrations:
nveil install-skill
nveil mcp
Why NVEIL?
| Capability | NVEIL | Chatbot data analysis¹ | LLM-to-viz libraries² | Traditional plotting³ |
|---|
| Natural-language input | ✓ | ✓ | ✓ | ✗ |
| Raw data stays on your machine | ✓ | ✗ | ✗ | ✓ |
| Only schema + stats sent to server | ✓ | ✗ | ✗ | N/A |
| Deterministic, reproducible output | ✓ | ✗ | ✗ | ✓ |
| Offline re-rendering, zero API calls | ✓ | ✗ | ✗ | ✓ |
Portable saved specs (.nveil files) | ✓ | ✗ | ✗ | ✗ |
| 2D + 3D + geospatial + scientific | ✓ | 2D | 2D | varies |
| Multi-backend (Plotly, VTK, DeckGL) | ✓ | ✗ | ✗ | ✗ |
| Data processing engine | ✓ | ✓ | partial | ✗ |
¹ ChatGPT Advanced Data Analysis, Claude Analysis tool, Gemini Data Agent · ² PandasAI, LIDA, Julius, Vanna · ³ Plotly, Matplotlib, Seaborn
How It Works
Your Data ──> Toolkit ──metadata only──> NVEIL AI ──> Processing Plan ──> Local Execution ──> Result
^ ^
raw data stays here raw data stays here
- You describe what you want in plain language
- NVEIL AI plans the data processing and visualization (only metadata is sent — column names, types, statistics)
- The Toolkit executes locally — joins, aggregations, pivots, rendering — all on your machine
- You get a figure — Plotly, VTK, or DeckGL, auto-selected for your data
Key Features
🧠 Two Engines in One
Data processing (joins, pivots, aggregations, geocoding, time series) AND visualization generation from a single prompt.
🔒 Data Privacy by Design
Raw data never leaves your machine. Only column names, types, and aggregate statistics are sent.
📈 Multi-Backend Rendering
Auto-detects the best engine: Plotly (2D charts), VTK (3D/medical), DeckGL (geospatial).
|
🧪 Auditable Results
Powered by constraint solving, not random generation. Same input = same output, every time.
⚡ Offline Rendering
spec.render() runs 100% locally with zero API calls.
💾 Reusable Specs
Save to .nveil files, reload later, render on new data — no server needed.
|
Beyond Simple Charts
NVEIL handles geospatial heatmaps, 3D volumes, scientific visualizations, medical imaging (DICOM), biosignal data (EDF/EDF+), network graphs, and 50+ other visualization types — all from natural language.
Save Once, Render Forever
spec = nveil.generate_spec("Monthly trend by category", df)
spec.save("trend.nveil")
spec = nveil.load_spec("trend.nveil")
fig = spec.render(fresh_data)
nveil.save_image(fig, "report.png")
Installation
Requirements: Python 3.10+
Getting Started
- Create an account at app.nveil.com
- Generate an API key in Settings
- Start visualizing
import os
import nveil
nveil.configure(api_key=os.environ["NVEIL_API_KEY"])
spec = nveil.generate_spec("scatter plot of price vs area", df)
fig = spec.render(df)
nveil.show(fig)
See the examples/ directory for more usage patterns.
Documentation
Full documentation is available at docs.nveil.com:
Contributing
Contributions are welcome under the project's Contributor License Agreement. Bug reports and feature requests are welcome via GitHub Issues.
License
GNU AGPL v3 or later. See LICENSE. Commercial dual-licensing is available — contact pierre.jacquet@nveil.com.
Website •
Documentation •
Platform