Ship Governed AI Agents
Add runtime governance to any AI agent in under five minutes. Python & JavaScript SDKs, REST API, and framework-specific wrappers.
Five-Minute Quickstart
From zero to governed agent in four steps.
Create a Fleet
Sign up and create your first governance fleet. This gives you an API key and a workspace for your agents.
Create free account →Install the SDK
Add the Prüfer SDK to your project. Available for Python (pip) and JavaScript (npm).
Register an Agent
Register your AI agent with your fleet. This creates a governance identity for runtime enforcement.
Enforce Governance
Wrap agent actions with governance checks. Prüfer evaluates each action against your policy rules in real time.
from prufer import PruferClient
client = PruferClient(
api_key="fleet_YOUR_KEY",
base_url="https://api.prufer.ai/api/v1"
)
# Check governance before any agent action
decision = client.check_action(
agent_id="clinical-advisor",
action="recommend_treatment",
context={
"patient_age": 72,
"diagnosis": "hypertension",
"proposed": "ACE inhibitor"
}
)
if decision.allowed:
# Proceed with the action
result = agent.run(action)
else:
# Handle block or escalation
print(f"Blocked: {decision.reason}")SDKs & Libraries
Choose your language. Every SDK has feature parity.
Python SDK
Native support for LangChain, CrewAI, OpenAI Assistants, AutoGen, and a universal decorator pattern.
pip install prufer⚡JavaScript SDK
Governed wrappers for Node.js agents. Drop-in middleware for LangChain.js, Vercel AI SDK, CrewAI, and custom frameworks.
npm i @prufer/sdk🔌REST API
Language-agnostic HTTP endpoints for governance checks, agent registration, audit logs, and trust scores.
OpenAPIFramework Integrations
Drop-in governance for the most popular agent frameworks. Zero architecture changes.
LangChain
Python & JS callbacks
Vercel AI
Tool wrapper
CrewAI
Step & task hooks
OpenAI
Assistants API
AutoGen
AG2 hooks
HTTP Proxy
Any LLM call
Decorator
Universal wrap
Agent Registration
# Register an agent with your fleet
curl -X POST https://api.prufer.ai/v1/enforce/register \
-H "X-Fleet-Key: fleet_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"agent_id": "clinical-advisor",
"name": "Clinical Decision Support",
"type": "healthcare",
"metadata": {
"model": "gpt-4o",
"department": "cardiology"
}
}'API Reference
Core endpoints for runtime governance.
/api/v1/enforce/checkGovernance Check
Evaluate an agent action against fleet policies in real time. Returns allow, block, or escalate.
/api/v1/enforce/registerAgent Registration
Register agents with your fleet. Supports batch registration and custom metadata.
/api/v1/enforce/logsAudit Logs
Query the decision log for governance events. Filter by agent, action, date, or decision type.
/api/v1/enforce/escalationsEscalations
Retrieve and manage human-in-the-loop escalations. Approve, reject, or reassign.
/api/v1/trust/agents/:idTrust Scores
Retrieve trust scores and behavioral history for any registered agent.
Developer Resources
Ready to integrate?
Create a free fleet, get your API key, and start making governance checks in under five minutes.