Lyzr-Cognis/cognis
Lightweight, local-first memory for AI agents. Hybrid vector + BM25 search, LLM-powered fact extraction, zero infrastructure — just pip install.
This project helps AI agent builders give their agents persistent memory. It takes natural language conversations and extracts key facts, allowing agents to recall past interactions and information. The output is a searchable, categorized memory bank that makes agents more consistent and helpful. It's for developers creating AI agents, chatbots, or intelligent assistants that need to remember details about users or ongoing conversations.
Use this if you are building AI agents and need a simple, local way for them to remember details and context from conversations without managing complex infrastructure.
Not ideal if you need a shared, large-scale memory solution for many agents across multiple servers, or if you prefer a fully managed cloud service.
Stars
37
Forks
—
Language
Python
License
MIT
Category
Last pushed
Mar 27, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/vector-db/Lyzr-Cognis/cognis"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
memgraph/memgraph
Open-source graph database, tuned for dynamic analytics environments. Easy to adopt, scale and own.
arpitnath/blink-query
DNS-inspired knowledge resolution layer for AI agents. Typed records, FTS5 search, query DSL,...
timelabs-npo/rhea-memory
Python memory layer — SQLite KV store, timeline, compact context generator. pip install rhea-memory.
cpdata/meshmind
MeshMind is a knowledge management system that uses LLMs and graph databases to store and...
longevitycoach/StrunzKnowledge
This project aims to create a comprehensive, searchable knowledge base from the online articles...