Development · Web app
AIMap
Discovery and security testing platform for exposed internet-facing AI endpoints.
AIMap is an open-source reconnaissance and attack surface mapping tool created by Bishop Fox to find exposed AI and LLM infrastructure across the web. It uses Shodan queries and custom Nuclei templates to detect unauthenticated endpoints like Ollama, MCP servers, LangServe, and Gradio, scoring their security risk from 0 to 10. The platform visualizes discoveries on a 3D globe and includes automated testing suites for prompt injection, tool abuse, and data extraction.
What it's for
- Discover publicly accessible AI servers, LLM APIs, and MCP endpoints across company network CIDRs
- Identify unauthenticated Ollama, Gradio, or ComfyUI instances exposing sensitive models or code execution tools
- Run automated security testing against AI endpoints to check for prompt leakage and tool invocation vulnerabilities
- Monitor global exposure of AI infrastructure on an interactive 3D map
Features
- Shodan integration with 32+ curated queries targeting AI protocols and frameworks
- Custom Nuclei templates to fingerprint MCP servers, LangServe, and OpenAI-compatible APIs
- 0 to 10 automated risk scoring based on authentication status, exposed tools, CORS, and TLS
- Protocol-specific attack suites for MCP tool abuse, Ollama extraction, and prompt injection
- Interactive 3D globe interface powered by globe.gl with filter chips and real-time WebSocket feeds
- Shodan-style search syntax allowing queries by risk, protocol, country, exposed tool, or leaked prompt
- Optional serverless workload offloading to Modal containers
- Docker Compose deployment bundling FastAPI, MongoDB, Redis, and a React frontend
Pricing
Open source. Free and open source (requires own Shodan API key)
Details
- Type
- Web app
- Category
- Development
- Topic
- Security & Pentesting
- Open source
- Yes
- Platforms
- Self-hosted, Web, Command line
- For
- Security researchers, penetration testers, and DevSecOps teams securing AI deployments.
- Alternative to
- Shodan, Censys, ProjectDiscovery Nuclei
- Tags
- security, reconnaissance, osint, llm, mcp, shodan, vulnerability scanner
Why it was saved
- @carterperez.dev: Enables security researchers to identify publicly exposed AI servers (such as Ollama, ComfyUI, or Gradio instances) with zero authentication, scoring their risk levels.
Related
Check if an email address is registered across 120+ online platforms.
AI-powered dark web OSINT tool for automated querying and summaries
Turn websites and documents into clean, LLM-ready markdown or structured data.
Single API and playground for hundreds of multimodal AI models
Python interactive CLI toolkit for automated OSINT, network reconnaissance, and security checks
Real-time GitHub monitoring tool to find leaked credentials and API tokens
Updated