metronix-memorySelf-hosted agent memory stack in Docker: Postgres + Qdrant + Neo4j hybrid retrieval, a temporal knowledge graph, ontology layer and freshness checks behind one MCP-native API.
Why switchSame job — self-hostable agent memory benchmarked on LongMemEval. Hindsight bets on reflect-and-learn over Postgres; Metronix bets on retrieval breadth (dense + sparse + graph). Pick by whether your failure mode is recall or reasoning over what was recalled.
Full comparison → MemMachineLong-term memory layer for AI agents — episodic (graph), profile (SQL) and working memory behind Python/TS SDKs, REST and MCP; ships LangChain, LangGraph, CrewAI and LlamaIndex integrations.
Why switchBoth self-hostable long-term memory services for agents behind Python/TS SDKs and REST. MemMachine splits episodic/profile/working memory with framework integrations; Hindsight bets on retain/recall/reflect and benchmark-topping learned memory.
Full comparison → agent-memoryNeo4j Labs' graph-native agent memory: conversations, a POLE+O entity knowledge graph and reasoning traces in one store, with a 16-tool MCP server and hosted or self-hosted backends.
Why switchSame reflect-and-learn premise, different substrate: Hindsight runs on Postgres and benchmarks on LongMemEval; this trades that for a real graph with entity resolution and audit edges. Choose by whether you'd rather write Cypher or keep one Postgres dependency.
Full comparison → agentic-context-engineLearning loop for any agent: reflect on failures, distill strategies into a Skillbook, inject them next run — 2x consistency on Tau2, 49% token cuts. LiteLLM-based, 100+ providers.
Why switchBoth sell 'agents that learn, not just remember' — opposite deployments: Hindsight is a self-hosted memory service (Docker + Postgres, retain/recall/reflect over HTTP); ACE is an embedded Python loop distilling strategies in-process. Service vs library.
Full comparison → LightMemICLR 2026 memory framework for LLMs/agents: LLMLingua pre-compression, topic segmentation and offline memory updates — leading LoCoMo/LongMemEval results at lower token cost.
Why switchBoth chase LongMemEval leadership from opposite ends: LightMem is a research framework optimizing token cost via pre-compression; Hindsight is a packaged service optimizing accuracy. Benchmark rivals, different deployment realities.
Full comparison → AcontextSkill memory layer for agents: auto-captures learnings from runs into plain Markdown skill files you can read, edit, git and share across frameworks — memory without an opaque store.
Why switchTwo shapes for 'memory that learns from its runs': hindsight is a service your app calls (retain/recall/reflect over Postgres), acontext an artifact your agent reads (Markdown skill files you can git and edit). Decide database or repo file before you build on either.
Full comparison → MemoriaRust memory layer for AI agents with Git-style version control — snapshot, branch, merge and rollback over MatrixOne's copy-on-write engine, plus hybrid vector + full-text retrieval.
Why switchBoth add governance to agent memory; hindsight is retain/recall/reflect over Postgres, Memoria is version-controlled 'Git for memory' with auditable snapshots.
Full comparison →