MemTensor/MemT-1.0

Metis is the memory foundation model series developed by Metis Group, MemTensor.

The official repo of the paper: $delta$-mem: Efficient Online Memory for Large Language Models

A language model that forms persistent memories from conversation and maintains them through sleep. MEMIT weight editing + null-space-constrained maintenance.

Plug-and-play memory for LLMs in 3 lines of code. Add persistent, intelligent, human-like memory and recall to any model in minutes.

MEME: Multi-Entity & Evolving Memory Evaluation — reference implementation (companion to arXiv preprint)

Hybrid Transformer with Online RNN Memory for Neocortex Memory.

Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.

Memory layer for AI agents — SDKs, CLI and MCP server for Membase

Drop-in AI memory layer with 2x faster retrieval and 10x lower cost. Fully compatible with Mem0 API. Migrate in 5 minutes without any code changes. Self-host for free.