Development moved: This integration is maintained in the DeepKeep integrations monorepo. Please open new issues and pull requests in the monorepo. Historical source and existing links remain available here. The PyPI package name remains
crewai-tools-deepkeep.
DeepKeep AI Firewall integration for CrewAI native guardrails and execution hooks.
This package is designed as an external CrewAI integration, not a patch to crewai_tools. It follows CrewAI's native extension points:
Task.guardrail/Task.guardrailsfor task output validation and retry feedback.- Execution hooks for pre/post LLM and tool-call enforcement.
- Optional CrewAI tools for agent-driven moderation workflows.
pip install crewai-tools-deepkeepSet credentials:
export DEEPKEEP_API_KEY="dk_..."
export DEEPKEEP_BASE_URL="https://api.example.deepkeep.ai"Use this when you want CrewAI to validate task output and retry the task when DeepKeep blocks it.
from crewai import Agent, Task
from crewai_deepkeep import DeepKeepGuardrail
deepkeep = DeepKeepGuardrail(
post_model="output-firewall-id",
)
task = Task(
description="Answer the user safely.",
expected_output="A safe answer.",
agent=agent,
guardrail=deepkeep.check_output,
guardrail_max_retries=3,
)DeepKeepGuardrail.check_output returns CrewAI's expected guardrail tuple:
(True, validated_or_modified_output)
(False, "feedback for the agent retry")Use this when you want DeepKeep to enforce policy around every model and tool boundary.
from crewai_deepkeep import DeepKeepHooks, install_deepkeep_hooks
install_deepkeep_hooks(
DeepKeepHooks(
pre_model="input-firewall-id",
post_model="output-firewall-id",
)
)This registers checks for:
PRE_MODEL_CALLPOST_MODEL_CALLPRE_TOOL_CALLPOST_TOOL_CALL
Blocking decisions raise CrewAI HookAborted with source="deepkeep". Redact/modify decisions replace the intercepted message, response, tool input, or tool result where CrewAI supports replacement.
If you prefer explicit registration:
from crewai.hooks import InterceptionPoint, on
from crewai_deepkeep import DeepKeepHooks
hooks = DeepKeepHooks(pre_model="input-firewall-id", post_model="output-firewall-id")
on(InterceptionPoint.PRE_MODEL_CALL)(hooks.check_model_input)
on(InterceptionPoint.POST_MODEL_CALL)(hooks.check_model_output)
on(InterceptionPoint.PRE_TOOL_CALL)(hooks.check_tool_input)
on(InterceptionPoint.POST_TOOL_CALL)(hooks.check_tool_output)Tools are available for workflows where the agent should explicitly call DeepKeep:
from crewai_deepkeep.tools import DeepKeepModerateInputTool
agent = Agent(
role="AI Security Guard",
tools=[DeepKeepModerateInputTool(model="input-firewall-id")],
)For production enforcement, prefer task guardrails and hooks over agent-callable tools.
If submitting this to CrewAI upstream, position it as an external integration page, not as a request to add DeepKeep classes into CrewAI core:
DeepKeep integrates with CrewAI's native task guardrails and execution hooks. It is distributed as an external package, so CrewAI core does not need to maintain vendor-specific API code.
DEEPKEEP_API_KEY: DeepKeep API key.DEEPKEEP_BASE_URL: DeepKeep API base URL, without a trailing slash.
Current integration directory:
https://github.com/Deepkeepai/deepkeep-integrations/tree/main/integrations/crewai