Awesome Memory Papers in Vision-Language Models

LLM-agnostic memory layer for AI agents. No embeddings, no vector DB — just fast, structured, temporal memory that any LLM can consume as plain text.

An AI partner with true active memory. She keeps a diary in her own voice, digests her days in nightly dreams, knows what today is, and comes back tomorrow remembering the life you share. Voice, desktop avatar (Live2D / VRM / three.js / html vanilla), a multi-layered self. Companionship is one use case; the memory is the point. Runs on Claude Code.

Persistent causal memory for AI agents. 295x faster than Mem0. LangChain, LlamaIndex, AutoGen, CrewAI. Rust, zero deps.

Self-hosted memory + evidence layer for AI agents (Claude Code, Codex, Hermes, ...) — embeddable Go library, MCP / HTTP / CLI, evidence-backed claims, bitemporal recall, axi-go execution kernel with JSONL audit + token budgets, cosign-signed releases with SLSA L3 provenance. No vendor cloud, no per-call billing.

A lightweight, pluggable memory backend for agent-based simulations. Supports temporal data, experience replay, and persistent state logging

Synapse Agent Memory — local-first long-term memory for coding agents: temporal truth, cited context, feedback learning, and verified self-healing in private SQLite.

Durable, private, time-aware memory engine for long-running AI agents

Local-first memory kernel for coding agents: MCP-native, temporal, auditable, source-keyed, and benchmarked.

Sellmind - Temporal Memory for AI. Cross-session memory persistence with emotional coherence. Built for Claude Code agents.

Time-aware memory system that understands when things happened, not just what happened

Offline-first deterministic agent harness with GraphRAG, temporal memory, DoT DAG planning, and SLM/LLM routing.