ActiveFence is now Alice
x
Parlant Integration

Maintain control with Parlant and Alice AI guardrails.

Agent-based systems built on Parlant require policy enforcement and runtime safeguards in production. Deploy Parlant agents with Alice-enforced guardrails, continuous testing, and risk visibility to prevent abuse and maintain operational control.

Code snippet showing Guardrails setup from wonderfrence and parlamt agents with policies and enforcement mode.
How it works

INTERCEPT USER PROMPTS BEFORE EXECUTION
MONITOR EVERY AGENT ACTION AND OUTPUT
INTERCEPT USER PROMPTS BEFORE EXECUTION
parlant_agents.py
from wonderfence import Guardrails
from parlant.agent_guardrails import Agent

guardrails = Guardrails(
    policies=["prompt_injection", "pii_detection", "tool_misuse"],
    mode="enforce",
    log_violations=True
)

agent = Agent("customer-support-parlant")

response = guardrails.run(
    agent=agent,
    input=user_message,
    context={"user_role": "customer", "channel": "web_chat"}
)

return response.safe_output
Stay on your toes

What WonderFence adds to Parlant

Parlant gives you the agent. WonderFence gives you the controls — dynamic guardrails, granular policies, and full runtime visibility on top of what Parlant ships by default.

Policies That Match Your Business

Parlant's default guardrails are generic. WonderFence lets you enforce policies that reflect your compliance rules, brand voice, and risk tolerance — preconfigured options out of the box, or write your own. Apply them organization-wide or per-agent.

Defense Against Prompt Injection and Tool Misuse

Stop jailbreaks, data exfiltration, and tool abuse before they hit your databases, APIs, or customers. WonderFence catches malicious patterns that get past Parlant's default guardrails.

Visibility Into What Agents Actually Do

Parlant shows you what you built. WonderFence shows you what your agents are doing in production — every prompt, tool call, and output, plus drift, policy violations, and unexpected behavior as it happens.
Trusted by the world’s biggest enterprises & foundation model labs