Nuclei API provides a REST API to manage vulnerability detection templates, generate AI-assisted Nuclei templates, validate and refine them, and run scans against targets. It orchestrates an asynchronous pipeline using Celery with Redis as the broker and cache, and leverages Ollama for LLM-driven template generation and refinement. You can interact with the API to fetch vulnerabilities, generate AI-generated templates, upload your own YAML templates, and trigger scans on IPs or domains. The MCP integration exposes tool-like endpoints that allow an LLM agent to call into the Nuclei scanning and template management capabilities, enabling AI-powered security workflows. The platform runs inside Docker containers and can be accessed through its API surface or the provided frontend when deployed with the accompanying stack.

Prerequisites

• Docker and Docker Compose
• Python 3.8+ (for development or backend scripts) and Node.js 18+ (for frontend if needed)
• Redis (or Dockerized Redis as part of the stack)
• Ollama (or a compatible LLM API) for template generation

Step-by-step installation

1. Install dependencies and clone the repository

   git clone <repository-url> cd nuclei-api

2. Install and start the stack with Docker Compose (recommended for production-like setup)

   docker compose up -d

3. Optional: Run locally for development (fast bootstrap)

   Start Redis and Ollama containers (as needed by the app)

   docker compose -f docker-compose.yml up -d redis ollama

   Run the API using uvicorn (if developing locally)

   uvicorn app.main:app --reload --host 0.0.0.0 --port 8080

4. Access the API and frontend

   • API: http://localhost:8000 (Docker deployment) or http://localhost:8080 (local uvicorn)
   • Frontend (if enabled): http://localhost:3000

Tips and common issues:

• Ensure Redis is reachable at the configured URL (default redis://localhost:6379/0) since Celery relies on Redis for broker and results backend.
• Ollama must be running and accessible to generate/refine templates; verify the Ollama endpoint and model availability.
• When using Docker-compose, you can scale celery workers as needed with docker compose up -d --scale celery_worker=3 to handle higher loads.
• For MCP integration, expose the tool manifest and ensure LLM agents can discover and call the available nuclei-api routes (NucleiRoutes, PipelineRoutes, MCP Routes).
• If you’re implementing a custom frontend, the API exposes endpoints for template generation, storage, and scans; review the API structure under app/api for exact routes and payloads.