I built a better, human like memory, for Agents
emson
11 points
1 comment
March 29, 2026
Related Discussions
Found 5 related stories in 725.0ms across 14,015 title embeddings via pgvector HNSW
- We built a persistent agent memory layer on Elasticsearch with 0.89 recall showmypost · 103 pts · June 18, 2026 · 61% similar
- Agent Memory: An Anatomy brgsk · 38 pts · May 27, 2026 · 61% similar
- Memanto; open-source memory agent that remembers, recalls and answers majidfekri · 14 pts · June 18, 2026 · 58% similar
- MenteDB – open-source memory database for AI agents (Rust) mentedb · 15 pts · April 24, 2026 · 58% similar
- Show HN: AI memory with biological decay (52% recall) SachitRafa · 82 pts · April 26, 2026 · 58% similar
Discussion Highlights (1 comments)
slowmovintarget
Love the idea. Please protect your supply-chain. Python projects in the wild, especially as relates to agents, give me the willies these days with the GitHub Actions compromise that's been rolling through the OpenClaw dependencies. (Last one I saw was Telnyx.) How does this blend with something like wedow/ticket, or beads? Those are "short term" in-repo task lists that help shape context. I've been thinking that adding a good long-term memory was the next step. I'll have to kick the tires tomorrow.