NVEIL — Community Edition

Chat with your data — explore, process, and visualize it, fast.
The self-hosted, open-source edition of NVEIL.

Quick start • Why NVEIL • From code • Community • Contributing • License

---

NVEIL is the fast, conversational way to explore, process, and visualize data — any kind, any domain. Point it at your data, say what you want in plain language, and NVEIL prepares it and turns it into interactive, production-ready visualizations. No plotting code, no dashboards to wire by hand.

This is the Community Edition — the complete platform, free and open under the AGPL, that you self-host. Don't want to run your own server? The same platform is available hosted at nveil.com — a free tier plus paid plans, nothing to install.

▶  Try the live demo  —  no install, free to start.

Quick start

Self-host in minutes with prebuilt images — nothing to build.

Prerequisites: Docker + Docker Compose, and one LLM provider — a Gemini / OpenAI / Anthropic / Mistral key, or a local model (Ollama, llama.cpp, any OpenAI-compatible endpoint).

1. Download docker-compose.yaml into an empty folder — it's the only file you need.

2. Run the setup wizard, then open http://localhost:3000 to configure (it writes your .env):

docker compose up setup

3. Pull the images and launch:

docker compose up -d

Open https://localhost:8000 — and start chatting with your data. 🎉

Developer setup

Want to debug, turn on LLM tracing, or build the images yourself? Clone the repo and use docker-compose.dev.yml — it builds from source, with live reload and TEST mode (debug routes):

git clone https://github.com/nveil-ai/nveil.git
cd nveil

docker compose -f docker-compose.dev.yml up setup                       # configure → localhost:3000
docker compose -f docker-compose.dev.yml --profile core up --build -d   # build & run → localhost:8000

Optional LLM tracing with the bundled Langfuse (→ http://localhost:3030):

docker compose -f docker-compose.dev.yml --profile core --profile tracing up -d

Why NVEIL?

• 💬 Conversational — describe what you want in plain language; no plotting code, no manual dashboards.
• 🧮 Explore and process — clean, transform, aggregate, and join your data, not just chart it.
• 🌐 Any data, any domain — from a quick CSV to scientific datasets: business analytics, geospatial, 3D, imaging, signals, networks, and beyond.
• 🔒 Private by design — raw data stays on your infrastructure; only metadata reaches the AI.
• 🎯 Deterministic — visualizations come from constraint solving, not guesswork: same request → same result, every time.
• 🤖 Bring your own model — Gemini, OpenAI, Anthropic, Mistral, or a local model.
• 🗂️ A real app — chat, dashboards, file management, multi-user rooms, internationalization.

Use it from your code

The web app is the main way to use NVEIL. To drive it programmatically — from a Python script, your terminal, or an AI agent — there's an optional client, the NVEIL Toolkit:

pip install nveil

A Python SDK, a nveil CLI, and an MCP server for agents (Claude Code, Cursor, …), all pointed at your NVEIL instance. See nveil-toolkit.

Community

Questions, ideas, or want to show what you built? Join us:

• 💬 Discord — chat with the team and other users.
• 🗣️ GitHub Discussions — Q&A and proposals.
• 🐛 Issues — bug reports and feature requests.

Contributing

Contributions are welcome — the developer setup above is all you need to get started.

Pull requests are accepted under the project's Contributor License Agreement — you sign it once, on your first PR, by posting a one-line comment (a bot walks you through it). The CLA is a license grant, not an assignment: you keep ownership of your contribution and your moral rights.

If NVEIL is useful to you, a ⭐ helps others discover it.

Security

Found a vulnerability? Please report it privately — see SECURITY.md. Don't open a public issue for security reports.

License

NVEIL is dual-licensed:

• Open source — GNU AGPL-3.0-or-later: use, study, modify, and self-host freely. Note that AGPL §13 requires you to offer the corresponding source of any modified version you run as a network service. See LICENSE.
• Commercial — to embed NVEIL in a closed-source product, or run a modified hosted version without publishing your changes, a commercial license removes the copyleft obligations. See COMMERCIAL-LICENSE.md — contact pierre.jacquet@nveil.com.

Contributors

Thanks to everyone who helps make NVEIL better 💜

---

Website • Hosted platform • Documentation • Toolkit (SDK · CLI · MCP)