agentmem-core

agentmem-core

The core of AgentMem: a proactive memory layer for long-horizon coding agents. It maintains a structured memory bank and decides when to remind your agent, instead of dumping everything or retrieving on similarity.

pip install agentmem-core

Published as agentmem-core (the name agentmem was already taken on PyPI by an unrelated project); everything you import and run is still agentmem.

from agentmem import MemorySession, triggers

mem = MemorySession(task="Fix the auth tests without changing the public API")
while not done:
    reminder = mem.pending_context()          # str | None
    reply = call_your_agent(messages, memory_context=reminder)
    mem.observe(reply.new_messages)

Full docs, integrations, and the design rationale live in the main repository.

Licensed under Apache-2.0.