CREWAI · PYTHON

CrewAI Integration

Persistent, semantically-recalled memory for CrewAI agents. Your crews remember everything — across sessions, across restarts. Dakera handles embedding, storage, and retrieval server-side.

Package: crewai-dakera · GitHub →

Quick Start

Run Dakera

Dakera is a self-hosted memory server. Spin it up with Docker:

docker run -d \
  --name dakera \
  -p 3300:3300 \
  -e DAKERA_ROOT_API_KEY=dk-mykey \
  ghcr.io/dakera-ai/dakera:latest

# Verify it's running
curl http://localhost:3300/health

Install

# Core + integration
pip install crewai-dakera

# With CrewAI (if not already installed)
pip install "crewai-dakera[crewai]"

Requirements: Python ≥ 3.10, a running Dakera server.

Add memory to your crew

from crewai import Crew, Agent, Task
from crewai.memory import LongTermMemory
from crewai_dakera import DakeraStorage

storage = DakeraStorage(
    api_url="http://localhost:3300",
    api_key="dk-mykey",
    agent_id="my-crew",
)

crew = Crew(
    agents=[...],
    tasks=[...],
    memory=True,
    long_term_memory=LongTermMemory(storage=storage),
)

result = crew.kickoff(inputs={"topic": "AI trends"})

Your crew now persists everything it learns across runs.

Configuration

Parameter	Type	Default	Description
`api_url`	str	—	Dakera server URL (e.g. `http://localhost:3300`)
`api_key`	str	`""`	API key set via `DAKERA_ROOT_API_KEY`
`agent_id`	str	—	Logical identifier for this crew's memory namespace
`min_importance`	float	`0.0`	Minimum importance score for recalled memories
`top_k`	int	`5`	Number of memories to surface per turn

Using environment variables

import os
from crewai_dakera import DakeraStorage

storage = DakeraStorage(
    api_url=os.environ["DAKERA_URL"],
    api_key=os.environ["DAKERA_API_KEY"],
    agent_id="research-crew",
)

Full example — research crew

from crewai import Agent, Task, Crew, Process
from crewai.memory import LongTermMemory
from crewai_dakera import DakeraStorage

dakera = DakeraStorage(
    api_url="http://localhost:3300",
    api_key="dk-mykey",
    agent_id="research-crew",
)

researcher = Agent(
    role="Senior Researcher",
    goal="Uncover groundbreaking insights in {topic}",
    backstory="An expert researcher with decades of experience.",
    verbose=True,
)

writer = Agent(
    role="Content Writer",
    goal="Craft compelling reports based on research findings",
    backstory="A skilled writer who turns complex ideas into clear prose.",
    verbose=True,
)

research_task = Task(
    description="Research the latest developments in {topic}",
    expected_output="A detailed research report",
    agent=researcher,
)

write_task = Task(
    description="Write a blog post based on the research",
    expected_output="A polished 500-word article",
    agent=writer,
)

crew = Crew(
    agents=[researcher, writer],
    tasks=[research_task, write_task],
    process=Process.sequential,
    memory=True,
    long_term_memory=LongTermMemory(storage=dakera),
    verbose=True,
)

# First run — learns and stores findings
result = crew.kickoff(inputs={"topic": "quantum computing"})
print(result.raw)

# Second run — recalls prior research automatically
result = crew.kickoff(inputs={"topic": "quantum computing advances"})
print(result.raw)

How it works

After each task, CrewAI calls DakeraStorage.save() with the result
Dakera embeds the content server-side and stores it with a semantic vector
Before the next task, CrewAI calls DakeraStorage.search() — Dakera performs hybrid search (vector + BM25) and returns the most relevant past memories
Memories decay gracefully over time based on access patterns — frequently-accessed memories stay prominent