for AI Agents & MCP Servers
Cryptographic identity, capability-based access control, and continuous trust evaluation. SDKs for Python, TypeScript, and Java. Free for up to 25 agents.
$ npm install -g opena2a$ opena2a login
$ opena2a identity create --name my-agent --server cloud
# Agent registered with Ed25519 keypair
# Trust scoring active
# Audit trail logging
$ opena2a identity trust my-agent
# Trust score: 92/100 (8 factors evaluated)Every agent gets a unique Ed25519 keypair. Request signing, verification, and attestation built in. No shared credentials, no API keys to rotate.
Continuous behavioral analysis across 8 dimensions. Scores update in real time. Permissions tighten or relax automatically based on trust evaluation.
Automatic discovery of MCP servers your agents connect to. Shadow MCP detection, tool capability inventory, and connection monitoring.
Agents are restricted to explicitly declared capabilities. Runtime enforcement catches privilege escalation before it happens.
Every agent action logged with cryptographic attribution. Immutable records, SOC 2 ready export, full compliance support.
Manage agents, MCP servers, trust scores, and security policies from one dashboard. Role-based access for your team. Alerts on anomalies.
SDKs for Python, TypeScript, and Java. Auto-detects LangChain, CrewAI, AutoGen, and other frameworks. Express and Fastify middleware included.
Python SDK
aim-sdk
PyPI
TypeScript SDK
@opena2a/aim-sdk
npm
Java SDK
org.opena2a:aim-sdk
Maven
CLI
opena2a
npm
Free tier: 25 agents, 25 MCP servers, 5 team members, full audit trail