Tessera
A Python SDK that gives any LLM agent long-term memory: it extracts facts and episodes from chat turns, handles corrections and forgetting, and returns deterministic ranked recall, with a drop-in MCP server for Claude Code, Cursor, and Codex.
I've tried most of the agent-memory libraries out there, and they all felt thin in the same two places: temporal knowledge (when things happened, what's since changed) and spatial knowledge (how facts relate to each other). Tessera is my attempt to give those a hand, not just store embeddings and hope.
It extracts facts and episodes from conversation turns, reconciles corrections and forgetting instead of piling up contradictions, and returns ranked recall that's deterministic: the same query gives you the same memories, which matters more than people admit when you're debugging an agent.
It ships as a sync + async Python SDK on PyPI and as a drop-in MCP server for Claude Code, Cursor, and Codex. I'm building it out and running evals to benchmark it against the top memory providers. I'd rather know where it loses than pretend it doesn't.