mcp server
Neb
NEB — Visual AI Agent Builder. Craft, customize and run AI agents on an interactive node canvas. Wire up third-party APIs, draw from libraries of skills, prompts and templates, and test agents live in
Set up
Runs on your computer. Pick your agent and operating system.
This project does not publish an install method we can turn into a config. Follow the README instead.
README
synced Oct 27, 2026About
NEB is a visual AI agent builder. Instead of wiring agents together with code, you compose them on an interactive node canvas: drag in building blocks, connect them into workflows, attach real APIs, pick from a library of skills, prompts and templates, and then test everything in a live, real-time execution playground.
App Preview
Features
- Visual node canvas — drag, connect and configure agent workflow nodes on an interactive flow canvas.
- Live execution playground — run agents in real time and watch output stream back as it happens.
- API integrations — hook up third-party APIs with preset connectors or define your own endpoints.
- Skill library — a large set of built-in skills plus agent templates to kick-start any project.
- Prompt library — store, organize and reuse prompts across agents.
- Template library — start from pre-built agent workflows instead of a blank canvas.
- Node configuration panel — fine-tune every node's inputs, outputs and behavior in one place.
- Table & prompt editors — full-featured inline editors for structured data and prompts.
- Embed & export — embed your agents and export/share artifacts, including Office docs via inline viewer.
- Custom Gemini agents — powered by the Gemini API with multi-provider AI support.
Getting Started
Prerequisites: Node.js 18+ and a Gemini API key.
# 1. Install dependencies
npm install
# 2. Create your environment file and add your API key
cp .env.example .env.local # then set GEMINI_API_KEY
# 3. Start the dev server
npm run dev
Open the printed local URL (default http://localhost:3000) to start building.
Scripts
| Command | Description |
|---|---|
npm run dev | Start the dev server (hot reload) |
npm run build | Build client + server production bundles |
npm run start | Run the built server |
npm run preview | Preview the built client |
npm run lint | Type-check the project with TypeScript |
Tech Stack
| Layer | Technologies |
|---|---|
| Frontend | React 19, TypeScript, Vite, Tailwind CSS 4, Motion (Framer Motion) |
| Backend | Node.js, Express, tsx |
| AI | AI SDK, @ai-sdk/google, @ai-sdk/openai, @google/genai |
| Extras | Composio (tool integrations), xlsx (spreadsheet handling) |
Environment Variables
| Variable | Required | Description |
|---|---|---|
GEMINI_API_KEY | Yes | Your Gemini API key |
Project Structure
neb/
├── src/ # Client source code
│ ├── components/ # UI: canvas, playground, config panels, editors
│ ├── data/ # Agents, skills, prompts, templates, providers
│ └── lib/ # Shared utilities
├── data/ # Runtime data (agents.json)
├── assets/ # Brand assets (logo, previews)
├── server.ts # Express server entry point
├── vite.config.ts # Vite configuration
├── tsconfig.json # TypeScript configuration
└── package.json
Contributing
- Fork the repository.
- Create a feature branch:
git checkout -b feat/my-feature - Commit your changes:
git commit -m "feat: add my feature" - Push to the branch:
git push origin feat/my-feature - Open a pull request.