Show HN: An open Add/Search interface for benchmarking agent memory

Category: ai-ml

Tags: benchmarking, agent-memory, evaluation, llm-agents, leaderboard

Score: 6.5/10 (Innovation: 7, Technical: 5, Documentation: 7, Utility: 7)

Agent Memory Leaderboard (AML) is an open evaluation platform and leaderboard for long-term memory systems in AI agents, standardizing how memory implementations are benchmarked via simple Add/Search interfaces. It offers a controlled, versioned evaluation protocol across textual and coding memory benchmarks, filling a real gap in comparing memory systems fairly.

Target audience: AI/ML researchers, LLM agent developers, memory system maintainers

Repository: https://twitter.com/AgentMemoryL/status/2101515447816663222 · Python · 1109 stars

View on Hacker News