External AI Surface Intelligence by Entelijan
UMAI continuously discovers, collects, and analyzes AI ecosystem data across the public internet to surface external AI surface intelligence. We maintain two core data sets: a public index of internet-wide AI data, and a corporate directory focused on exposure and change over time for AI-native companies.
Web App: umai.entelijan.com
UMAI is external AI surface intelligence for the enterprise. Monitor your corporate AI surface risk the way you monitor a credit score: continuously, from the outside, with the AI-specific depth that general-purpose scanners miss.
Where broad internet scanners catalog services generically, UMAI brings deep domain knowledge of the AI and ML ecosystem, spanning inference engines, model registries, vector databases, agent protocols, and the governance and compliance signals that companies publish about their AI. Our public index holds over 7 million published records and grows continuously.
UMAI maintains two core data sets. The categories below are representative, not an exhaustive list of everything we collect and discover.
A searchable index of internet-wide AI data:
- AI API and inference endpoints, including the raw response returned to a GET or POST request
- AI agent crypto transactions
- CVEs and exploits associated with specific AI models and AI applications
- AI services and their default ports
- Machine-readable manifests such as llms.txt, agents.md, and /.well-known/ucp
- Vendor-disclosed AI governance terms extracted from trust-center, legal, and compliance pages: model-training opt-out, model provenance and sub-processors, AI liability allocation, human-oversight commitments, and security or AI-management certifications
- Regulatory, litigation, and incident risk
A corporate directory tracking external AI exposure and change over time for AI-native companies and AI enablers:
- AI infrastructure: cloud AI services, routing and DNS footprint, and discovered subdomains
- Agentic threat exposure
- AI supply chain: published artifacts, dependency manifests, container base images, and model manifests, covering what a company runs, depends on, and pulls from
- Agentic surface: agentic protocols such as MCP and UCP, logic manifests, agents.md / skill.md / agents.json, capability context, and the knowledge layer of llms.txt and grounding docs an agent ingests
- AI CVEs and exploits
- Deployed technologies
- AI endpoint surface
- Legal and compliance posture: train-directives that guard corporate assets from unauthorized model ingestion, public regulatory actions, litigation and incident matters on record against the company's AI, and look-alike domains
UMAI runs proprietary agentic data pipelines that continuously discover, enrich, verify, and publish the external AI surface. Autonomous agents correlate signals across the public internet, classify findings against deep AI-specific knowledge, cross-check them for accuracy, and keep published records fresh as the surface changes, turning raw discovery into verified, structured intelligence.
Organizations can request removal of their assets from UMAI's public index. Email contact@entelijan.com with the IP addresses or CIDR ranges, and we will add them to our permanent exclusion list.
If you identify a critical exposure through UMAI, for example an LLM endpoint leaking sensitive context, please follow coordinated disclosure: do not access or modify anything beyond what is needed to confirm the exposure, report to the asset owner first, and notify us at contact@entelijan.com so we can flag the record appropriately.
Content in this repository is licensed under CC BY 4.0. You may share and adapt the material with attribution.
UMAI is a product of Entelijan. For enterprise inquiries, contact sales@entelijan.com.