Null-drift: A bare-metal O(1) Memory Backend for CrewAI (No VectorDB bloat)

General
1

Got tired of standard vectordbs eating up disk space and causing python bloat when running continuous local agents. so i gutted the db approach and built a headless rust daemon (null-drift).
​instead of storing exact text logs, it manages the agent’s memory as a continuous array using geometric decay. memory footprint stays completely flat at O(1) no matter how long the crew runs.
​i just wrote the python wrapper so it plugs natively into CrewAI as a custom StorageBackend. u can use it literally by changing one line in your config

from nulldrift_agents import NullDriftCrewStorage

crew = Crew(
    agents=[...], 
    tasks=[...], 
    memory=True, 
    # routes all memory to the rust daemon
  storage=NullDriftCrewStorage(gateway_url="http://127.0.0.1:8000", thread_id="production_crew")
)

the repo and wrapper code are here. if you are building local-first heavy systems and are hitting context limits, i’d love for u to test this out and see if u can break the rust daemon! null-drift