CrewAI Integration

Use Papr as shared memory across CrewAI tasks so agents can reuse prior findings and maintain continuity.

What You Build

Shared memory retrieval before task execution
Post-task writeback of outcomes and learnings
User and tenant scoped memory boundaries

Prerequisites

Papr API key in PAPR_MEMORY_API_KEY
CrewAI project with task lifecycle hooks
Stable per-user identifier mapped to external_user_id

Integration Pattern

Search Papr with the task objective.
Pass retrieved context to the active CrewAI agent.
Write task outcomes back to Papr memory.

Minimal Setup

Add pre-task retrieval to your crew execution flow.
Add post-task writeback after task completion.
Validate with repeated tasks for the same user scope.

Python Skeleton

import os
from papr_memory import Papr

client = Papr(x_api_key=os.environ.get("PAPR_MEMORY_API_KEY"))


def pre_task_context(task_prompt: str, actor_id: str):
    return client.memory.search(
        query=task_prompt,
        external_user_id=actor_id,
        enable_agentic_graph=True,
        max_memories=20,
        max_nodes=15,
    )


def post_task_learning(outcome: str, actor_id: str):
    return client.memory.add(
        content=outcome,
        external_user_id=actor_id,
        metadata={"role": "assistant", "category": "learning"},
        memory_policy={"mode": "auto"},
    )

Validation Checklist

Task startup includes retrieved Papr context
Task output is persisted and searchable
Shared memory improves multi-step workflows

Troubleshooting

If outcomes are not reused, verify post-task writes and check scope fields using Error Playbook.

Guides: Retrieval