Mnemo Cortex

•
161
•

Empowers AI agents with persistent, local, and cross-agent memory, facilitating deep recall, semantic search, and contextual awareness across sessions.

Mnemo Cortex is an open-source cognitive coprocessor providing AI agents with active, persistent memory. Built by a single maker, it offers features like auto-capture, smart classification, and overnight consolidation to ensure agents learn and remember effectively. It operates locally, without cloud dependencies or telemetry, offering deep recall, semantic search, and unique cross-agent dreaming capabilities for shared context. Designed to integrate with various LLM clients, from Claude and ChatGPT to local models, Mnemo Cortex ensures AI partners remember interactions and context, enhancing their long-term intelligence and utility.

主要功能

01Deep Recall with persistent memory across sessions and semantic search
02Structured Facts with key-value storage and confidence tracking
03147 GitHub stars
04Active Memory with auto-capture, smart classification, and overnight consolidation
05Cross-Agent Overnight Dreaming for shared context synthesis
06The Librarian for document discovery and workspace indexing

使用案例

01Integrating advanced memory capabilities with various LLM clients like Claude, ChatGPT, LM Studio, and Ollama.
02Providing persistent, shared memory for fleets of AI agents across multiple machines.
03Storing and recalling structured facts with confidence levels for entity attributes and settings.
04Enabling agents to find specific files and information within a local workspace via document discovery.