crewai-superlocalmemory

CrewAI storage backend powered by SuperLocalMemory V3 — local-first, privacy-compliant crew memory.

Features

• StorageBackend Protocol — Drop-in replacement for CrewAI's default LanceDB storage
• Local-first — All crew memories stay on your device (Mode A: zero cloud)
• EU AI Act compliant — Data never leaves the device, no DPA required
• Profile isolation — Different crews get isolated memory spaces

Installation

pip install crewai-superlocalmemory

Usage

from crewai import Crew, Agent, Task, Memory
from crewai_superlocalmemory import SuperLocalMemoryBackend

# Create SLM-backed storage
backend = SuperLocalMemoryBackend(profile_id="my-crew")

# Use with CrewAI Memory
crew = Crew(
    agents=[researcher, writer],
    tasks=[research_task, write_task],
    memory=Memory(storage=backend),
)

result = crew.kickoff()

Profile Isolation

# Each crew gets its own isolated memory space
alpha_backend = SuperLocalMemoryBackend(profile_id="crew-alpha")
beta_backend = SuperLocalMemoryBackend(profile_id="crew-beta")

Modes

Mode	Privacy	Use Case
a	Maximum (zero cloud)	EU AI Act compliant crews
b	High (local Ollama)	Private + LLM assist
c	Standard (cloud LLM)	Maximum accuracy

Links

• SuperLocalMemory
• Paper: arXiv:2603.14588
• GitHub

License

MIT

Part of Qualixar | Author: Varun Pratap Bhardwaj