CaptureNest

AI-powered self-hosted camera and media capture server

Capture → Analyze → Search — entirely on your own hardware

---

What is CaptureNest?

CaptureNest is a production-ready, self-hosted AI media server. Point it at your webcam or RTSP camera, capture photos and videos, and let a local AI model automatically describe and tag your media. Then search everything using natural language — no cloud, no subscriptions, no data leaving your machine.

"Show me when someone was at the front door"
"Find images with laptops"
"Show outdoor scenes from today"

---

Features

Feature	Description
Live Camera	Browser-based live preview, photo capture, and video recording
AI Analysis	Vision model (LLaVA) auto-generates descriptions and tags
Semantic Search	Natural language queries powered by vector embeddings + Qdrant
Media Library	Browse, filter, and manage your captures in a responsive grid
RTSP Support	Connect IP cameras and NVR/DVR systems via RTSP streams
Self-Hosted	Runs entirely locally — no cloud required
One-Command Deploy	Full Docker Compose setup

---

Architecture

┌──────────────────────────────────────────────────────────┐
│                   Web Dashboard (React)                  │
│   Dashboard │ Camera │ Library │ AI Search │ Settings    │
└─────────────────────┬────────────────────────────────────┘
                      │ HTTP (Vite proxy / nginx)
┌─────────────────────▼────────────────────────────────────┐
│              CaptureNest API (Fastify/TypeScript)        │
│                                                          │
│  ┌─────────────┐ ┌────────────┐ ┌────────────────────┐   │
│  │  Camera Svc │ │ Media Svc  │ │  AI Pipeline       │   │
│  │  (FFmpeg)   │ │  (sharp)   │ │  Ollama → Qdrant   │   │
│  └─────────────┘ └────────────┘ └────────────────────┘   │
│                                                          │
│  ┌──────────────────────────┐  ┌─────────────────────┐   │
│  │   SQLite (metadata)      │  │  Local Filesystem   │   │
│  └──────────────────────────┘  └─────────────────────┘   │
└──────────────────────────────────────────────────────────┘
         │                                      │
┌────────▼────────┐                   ┌─────────▼──────────┐
│  Ollama Server  │                   │  Qdrant Vector DB  │
│  (LLaVA model)  │                   │  (semantic search) │
└─────────────────┘                   └────────────────────┘

---

Quick Start

Prerequisites

• Docker and Docker Compose
• A webcam or RTSP camera (optional for initial setup)

1. Clone the repository

git clone https://github.com/yourdudeken/CaptureNest.git
cd CaptureNest

2. Start everything

docker compose up --build

3. Pull the AI vision model

In a separate terminal (first-time setup only):

# Pull vision model for image analysis
docker exec -it capturenest-ollama-1 ollama pull llava

# Pull embedding model for semantic search
docker exec -it capturenest-ollama-1 ollama pull nomic-embed-text

4. Open the dashboard

Navigate to http://localhost:3000

---

Local Development

Requirements

• Node.js 20+
• FFmpeg installed and in PATH
• A running Ollama instance (ollama serve)
• A running Qdrant instance (or use Docker: docker run -p 6333:6333 qdrant/qdrant)

Setup

# Install all dependencies
npm install

# Configure environment
cp server/.env.example server/.env
# Edit server/.env as needed

# Start both servers with hot reload
npm run dev

• API: http://localhost:4000
• Web: http://localhost:3000

---

Project Structure

CaptureNest/
├── server/                    # Backend (Node.js + TypeScript + Fastify)
│   ├── src/
│   │   ├── index.ts           # Entry point
│   │   ├── types.ts           # Shared type definitions
│   │   ├── db/
│   │   │   └── database.ts    # SQLite init + schema migrations
│   │   ├── api/routes/
│   │   │   ├── captureRoutes.ts   # Photo/video capture
│   │   │   ├── mediaRoutes.ts     # Media CRUD
│   │   │   ├── searchRoutes.ts    # AI search
│   │   │   ├── cameraRoutes.ts    # Camera management
│   │   │   └── settingsRoutes.ts  # App settings + health
│   │   └── services/
│   │       ├── ai/
│   │       │   ├── ollamaService.ts     # Vision model + embeddings
│   │       │   ├── qdrantService.ts     # Vector storage/search
│   │       │   └── analysisPipeline.ts  # Orchestration pipeline
│   │       ├── media/
│   │       │   ├── mediaService.ts     # File storage + DB
│   │       │   └── ffmpegService.ts    # Video recording
│   │       ├── camera/
│   │       │   └── cameraService.ts    # Camera config CRUD
│   │       └── settings/
│   │           └── settingsService.ts  # Config key-value store
│   ├── Dockerfile
│   └── package.json
│
├── client/                       # Frontend (React + Vite + Tailwind)
│   ├── src/
│   │   ├── main.tsx           # Entry point
│   │   ├── App.tsx            # Router
│   │   ├── lib/api.ts         # Typed API client
│   │   ├── components/
│   │   │   └── Layout.tsx     # Sidebar layout
│   │   └── pages/
│   │       ├── Dashboard.tsx  # Stats + recent captures
│   │       ├── Camera.tsx     # Live camera + capture
│   │       ├── Library.tsx    # Media grid browser
│   │       ├── MediaDetail.tsx # Full viewer + AI metadata
│   │       ├── Search.tsx     # Natural language search
│   │       └── Settings.tsx   # Configuration
│   ├── Dockerfile
│   └── package.json
│
├── docker-compose.yml         # Full deployment stack
├── LICENSE                    # MIT
└── README.md

---

Configuration

All settings can be changed in the Settings page of the dashboard, or via environment variables:

Variable	Default	Description
OLLAMA_URL	http://ollama:11434	Ollama server URL
OLLAMA_MODEL	llava	Vision model for image analysis
EMBED_MODEL	nomic-embed-text	Embedding model for search
QDRANT_URL	http://qdrant:6333	Qdrant server URL
MEDIA_PATH	./media	Where to store captured files
DB_PATH	./capturenest.db	SQLite database path
PORT	4000	API server port

---

API Reference

See docs/API.md for full endpoint documentation.

Quick overview:

POST /api/capture/image        – Capture still image
POST /api/capture/video/start  – Start video recording
POST /api/capture/video/stop   – Stop recording
GET  /api/media                – List media (paginated)
GET  /api/media/:id            – Get single media item
DELETE /api/media/:id          – Delete media
POST /api/media/:id/reanalyze  – Re-run AI analysis
POST /api/search               – Natural language search
GET  /api/settings/health      – Service health check

---

AI Models

CaptureNest uses Ollama to run AI models locally:

Model	Purpose	Pull command
llava	Image analysis, captioning, tagging	ollama pull llava
llava:13b	Higher quality analysis	ollama pull llava:13b
nomic-embed-text	Text embeddings for search	ollama pull nomic-embed-text

---

Docker Services

Service	Port	Description
capturenest-client	3000	React dashboard
capturenest-api	4000	Fastify API server
ollama	11434	AI model inference
qdrant	6333	Vector database

---

Contributing

Contributions are welcome! Please read CONTRIBUTING.md first.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

---

License

MIT — see LICENSE for details.