Distributed runtimes
Agents can operate on laptops, servers, IDEs, and short-lived local processes.
Available now · AI agent discovery
Know where supported AI agents operate.
Agent discovery establishes what exists and where. Interaction correlation is the next step toward understanding which tools and systems an agent used.
Direct definition
What is AI agent discovery?
AI agent discovery identifies supported agent software, runtimes, processes, and configuration evidence across managed endpoints.
AxLoop knowledge base
The visibility problem
Evidence must come before control.
Incomplete evidence
An installed agent does not prove it ran, and a running process does not prove which tool it invoked.
Changing behavior
Agent software and configurations can change as teams adopt new AI workflows.
How AxLoop helps today
Discover from the endpoint outward.
AxLoop identifies supported AI applications, agents, local models, runtimes, and MCP configuration evidence. Each finding keeps installed, configured, running, and independently verified states distinct.
- Identify supported software and configuration evidence
- Preserve device and user scope
- Monitor meaningful state changes
Questions teams ask
- Q1Where are supported AI agents installed?
- Q2Which agent processes were observed running?
- Q3What relationship evidence can AxLoop correlate?
Common questions
Answers, briefly.
- What counts as an AI agent in the enterprise?
- Any software that uses a model to plan and take actions, such as coding agents, desktop assistants with tools, automation scripts calling model APIs, and custom agents built on agent frameworks. Agents are defined by their ability to act, not only to answer.
- Why are agents harder to inventory than apps?
- Agents can be launched from a terminal, an IDE, or a script, may run only briefly, and gain capabilities through tools and MCP servers. A static software list misses that runtime behavior and the connections that give an agent its reach.
- What context matters for an agent finding?
- The device and user, agent runtime, model and MCP configuration evidence, timestamps, and evidence state all matter. Correlated or inferred relationships should remain distinct from direct observation.
- How does AxLoop discover agents?
- AxLoop observes supported agent software, processes, and configuration on managed devices. Findings are labeled by evidence state so installed, configured, and running are never confused.
Agent Interaction Observability