Integration

Add authorization to CrewAI agents

Give each CrewAI agent a warrant, and let it delegate a narrower one to the next — authority provably only shrinks down the crew.

Why authorize CrewAI tools?

CrewAI orchestrates multiple agents that hand work to each other. The multi-hop delegation is exactly where authority tends to leak: the second agent ends up with more power than intended. Warrants solve this by construction — a delegated warrant can only remove authority, never add it — so a sub-agent can never exceed what its delegator held.

Delegate down the crew, safely

The Python SDK verifies and attenuates warrants the same way:

from liceat import attenuate, verify_warrant, decode

# A researcher agent holds a read warrant; it delegates an even narrower one
# to a summarizer — the summarizer can never do more than read.
sub, _ = attenuate(researcher_warrant, atk, [{"type": "action", "actions": ["read"]}])

What you get

• Least privilege — each agent gets a warrant scoped to exactly the tools, resources, and spend it needs.
• Provable afterward — every allow/deny is recorded in a tamper-evident log you can audit or hand to a compliance reviewer.
• Instant revocation — kill a compromised agent's authority immediately, and it cascades to every sub-agent it delegated to.
• No new keys to leak — the agent never holds your API keys, only a narrow, expiring warrant.

Frequently asked

What is Liceat?

Liceat is permission infrastructure for AI agents. You issue a warrant — a scoped, spend-capped, revocable token — that grants an agent narrow authority; the agent presents it at the point of action; Liceat enforces it (scope, spend, rate, holder-binding, revocation) and records every decision in a tamper-evident log. It is open source with a zero-dependency SDK for TypeScript and Python.

Is it free?

The SDK and self-hosted gateway are free and open source. The hosted gateway at liceat.com gives 10,000 free decisions with no card, then meters at fractions of a cent per decision — payable by card or USDC.