AgentCMDB is the ServiceNow for AI Agents — a system of record, governance layer, and control plane for managing AI agents at enterprise scale.
Live Platform: Agent CMDB
Enterprises are rapidly deploying AI agents across workflows, tools, and platforms.
But there’s a growing problem:
- No central inventory of AI agents
- Ownership is unclear
- Governance is inconsistent
- Costs are invisible
- Failures are hard to trace
This creates “Agent Sprawl” — similar to shadow IT, but harder to control.
AgentCMDB provides:
- A CMDB for AI agents
- Full visibility across all agents
- Governance and policy enforcement
- Cost tracking and optimization
- A centralized AI control plane
Built for the era of AI agents — just as:
- ServiceNow manages IT assets
- Salesforce manages customer data
👉 AgentCMDB manages AI agents
- CIO / CTO → enterprise AI visibility & governance
- Platform teams → AI infrastructure & lifecycle
- Developers → agent registry & control
- Risk / compliance teams → guardrails & audit
- Inventory of all AI agents
- Ownership, metadata, dependencies
- Risk classification + lifecycle state
- Register agents from:
- LangChain / LangGraph
- CrewAI
- Copilot / APIs
- CI/CD pipelines
- View all agents across the organization
- Monitor workflows, failures, activity
- Track agent health and usage
- Dev → Test → Prod workflows
- Versioning + rollback
- Approval workflows (enterprise)
- Data access control
- Agent permissions
- Runtime guardrails
- Token usage tracking
- Cost per agent / workflow
- ROI visibility
A company has:
- 120 AI agents across finance, HR, and customer support
- Built using LangGraph, Copilot, APIs
With AgentCMDB:
✅ All agents are automatically registered
✅ Ownership and risk are defined
✅ Cost and usage are centrally tracked
✅ Policy violations trigger alerts
👉 Outcome:
- Governance
- Cost control
- Production readiness
AgentCMDB acts as a centralized control layer between AI agents and enterprise workflows.
AI Agents → AgentCMDB → Control Plane → Governance → Business Outcomes
Frontend (Control Tower UI)
│
▼
Agent Control Plane
- Lifecycle
- Governance
- Cost
- Observability
│
▼
Agent Registry (CMDB)
│
▼
Agent Runtime
(LangChain, CrewAI, APIs, Copilot)
- Python 3.10+
- Docker (optional)
git clone https://github.com/phaniburra-ai/agentcmdb
cd agentcmdb/backend
pip install -r requirements.txt
uvicorn app.main:app --reload
Open:
http://localhost:8000
docker compose up --build
from agentcmdb import register_agent
register_agent( name="Fraud Detection Agent", owner="Risk Team", framework="LangGraph", model_provider="Azure OpenAI", risk_level="critical" )
POST /api/agents
{ "name": "Invoice Processing Agent", "owner": "Finance Team", "environment": "prod", "risk_level": "high" }
AgentCMDB provides APIs to:
- Register and discover agents
- Track events and activity
- Attach governance policies
- Monitor cost and usage
- Agent registry (CMDB)
- Discovery API
- Basic control UI
- Governance + RBAC
- Lifecycle workflows (Dev → Test → Prod)
- Cost tracking and analytics
- Agent marketplace
- Policy enforcement engine
- Enterprise integrations (ServiceNow, Salesforce, Azure, etc.)
| Feature | OSS | Enterprise |
|---|---|---|
| Agent registry | ✅ | ✅ |
| API | ✅ | ✅ |
| Discovery | ✅ | ✅ |
| Basic UI | ✅ | ✅ |
| RBAC | ❌ | ✅ |
| Policy enforcement | ❌ | ✅ |
| SSO / Identity | ❌ | ✅ |
| Advanced analytics | ❌ | ✅ |
AgentCMDB aims to become the standard system of record and control plane for enterprise AI agents.
As organizations scale AI adoption, agents will become critical digital assets — requiring the same level of governance, visibility, and lifecycle management as:
- IT infrastructure (ServiceNow)
- Customer data platforms (Salesforce)
👉 AgentCMDB defines this new category for the AI era.
MIT License — see LICENSE for details. Enterprise features available under commercial license → Agent CMDB
Contributions welcome. Please read CONTRIBUTING.md or open an issue to discuss what you'd like to change.