CrewAI integration

crewai-dakera (0.3.0) gives CrewAI agents persistent semantic memory on a Dakera server. DakeraStorage saves and searches one agent's memories (one agent_id per instance); session, knowledge-graph, entity-extraction and namespace helpers are included. This page is for Python developers building crews with CrewAI.

Quick start

pip install crewai-dakera

Requires Python ≥ 3.10, dakera>=0.13.1 (installed automatically) and a running Dakera server. Works with Dakera server v0.12.0 and v0.11.108.

from crewai_dakera import DakeraStorage

storage = DakeraStorage(
    api_url="http://localhost:3000",
    api_key="dk-mykey",
    agent_id="crewai-researcher",
)

storage.save("User prefers executive summary format.")
results = storage.search("formatting preferences")

Features

DakeraStorage — persistent memory backend for CrewAI agents
DakeraSessionManager — session lifecycle for crew runs
DakeraKnowledgeGraph — entity-relationship operations
DakeraEntityExtractor — automatic entity extraction
DakeraNamespaceManager — multi-tenant namespace isolation

Examples

Basic agent memory

import os
from crewai_dakera import DakeraStorage

storage = DakeraStorage(
    api_url=os.environ.get("DAKERA_API_URL", "http://localhost:3000"),
    api_key=os.environ.get("DAKERA_API_KEY", ""),
    agent_id="crewai-researcher",
    search_k=3,
    importance=0.8,
)

storage.save("Completed market analysis: AI memory market growing 40% YoY.")
storage.save("Key competitor identified: Mem0 — open-source, Python-first.")

results = storage.search("market research findings")
for r in results:
    print(f"  [{r['score']:.3f}] {r['content']}")

Several agents, separate memories

from crewai_dakera import DakeraStorage

researcher = DakeraStorage(
    api_url="http://localhost:3000",
    agent_id="crewai-researcher",
    importance=0.8,
)

writer = DakeraStorage(
    api_url="http://localhost:3000",
    agent_id="crewai-writer",
    importance=0.8,
)

# Each agent stores to its own namespace
researcher.save("Python is the most popular language for AI/ML.")
writer.save("Blog outline: Top Languages for AI Development")

# Each agent recalls its own memories
results = researcher.search("AI languages")

API reference

DakeraStorage options

Parameter	Type	Default	Description
`api_url`	str	—	Dakera server URL
`api_key`	str	`""`	API key
`agent_id`	str	—	Agent identifier for memory namespacing
`search_k`	int	`5`	Results to return per search
`min_importance`	float	`0.0`	Minimum importance of returned memories
`importance`	float	`0.7`	Importance assigned to stored memories

Configuration

The package does not read environment variables. To keep the URL and key out of source code, read them yourself:

import os
from crewai_dakera import DakeraStorage

storage = DakeraStorage(
    api_url=os.environ["DAKERA_API_URL"],
    api_key=os.environ.get("DAKERA_API_KEY", ""),
    agent_id="crewai-researcher",
)