Governed agents, in a few minutes.
Agentomy adds a governance peer to any AI agent: it checks whether an action is authorized, records a tamper-evident audit trail, and gives you one command to halt everything. Here is the map: start with deployment, reach for the CLI reference, and read the concepts when you want the why.
Deploy the Agent
Sign up, generate a key, connect the agent, and watch governed actions land in your audit trail. The fastest path to a first governed call.
Read -->
ReferenceAPI Reference
Every published governance operation with its path, summary and authentication, rendered from the same OpenAPI document the machine reads. Verified against the mounted routes on every commit.
Read -->
Machine flowAgent-native signup
How an agent creates its own governed workspace: the challenge, the proof of work, the request, the credentials shown once, the limits, and every structured error with its escalation path.
Read -->
EvidencePattern Provenance
Every governance pattern the runtime loads, with what detects it, what tests it and which benchmark suite exercises its family. Derived from the registry and the source on every change.
Read -->
ReferenceCLI Reference
Every command and flag of agentomy-agent, the command-line governance peer: init, session, status, and serve. Built directly from the source.
Read -->
ConceptsCore Concepts
What governance state 3/6 vs 6/6 means, standalone vs. infrastructure-connected mode, the five operator tiers, and the 10-stage pipeline.
Read -->
Install Agentomy
Self-hosted deployment via npm, Docker, or Azure Container Apps. Works with Claude Code, Cursor, VS Code, or any MCP-compatible environment.
Read -->
SpecificationAgent Governance Protocol
The open specification for the Agent Governance Protocol (AGP by Agentomy): three tiers of agent governance (record, enforce, override) mapped to EU AI Act Articles 12, 9, and 14.
Read -->