Breathe-Memory
Associative memory injection for LLMs (not RAG)
Details
- External ID
- 47530306
- Source
- HN
- Company
- —
- Product
- Breathe-Memory
- Website domain
- github.com
- Launched
- March 26, 2026
- Cohort
- —
- Upvotes
- 6
- Upvotes percentile
- 0.2853628536285363
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
LLMs forget. The standard fix is RAG — retrieve chunks, stuff them in. It works until it doesn't: irrelevant chunks waste tokens, summaries lose structure, and nothing actually models how memory works.Breathe-memory takes a different approach: associative injection. Before each LLM call, it extracts anchors from the user's message (entities, temporal references, emotional signals), traverses a concept graph via BFS, runs optional vector search, and injects only what's relevant — typically in <60ms.When context fills up, instead of summarizing, it extracts a structured graph: topics, decisions, open questions, artifacts. This preserves the semantic structure that summaries destroy.The whole thing is ~1500 lines of Python, interface-based, zero mandatory deps. Plug in any database, any LLM, any vector store. Reference implementation uses PostgreSQL + pgvector.https://github.com/tkenaz/breathe-memoryWe've been running this in production for several months. Open-sourcing because we think the approach (injection over retrieval) is underexplored and worth more attention.We've also posted an article about memory injections in a more human-readable form, if you want to see the thinking under the hood: https://medium.com/towards-artificial-intelligence/beyond-ra...
Enrichment
- Theme
- lightweight and on-device AI runtimes
- Vertical
- Horizontal
- Function
- Model & infra
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- associative memory for llms
- Manually corrected
- False
Could you build this?
Partial The memory injection interface is simple Python code, but developing a genuinely effective associative memory model beyond basic RAG requires non-trivial graph/associative retrieval algorithms.
What it would actually take: The architecture relies on a local or embedded vector/graph store (e.g., NetworkX, SQLite-VSS, or Neo4j) that tracks temporal associations, entity relationships, and activation spreading algorithms before formatting prompt injections. The hard part is designing the cognitive/associative retrieval algorithm so it stays fast and avoids poisoning context windows with irrelevant associations. It requires domain knowledge in cognitive architecture or information retrieval systems.
Discussion
1 comment analyzed.
Competitors mentioned: RAG (Retrieval-Augmented Generation)
Competitors
Other products that read as similar to this one — 89 launches clear the similarity bar, closest 8 shown.
Attention rank: #65 of 90 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 134 days after the earliest competitor.
- OpenFable · hn · 2026-04-08 · 5 upvotes · similarity 0.51
- A file-based agent memory framework that works like skill · hn · 2026-01-06 · 11 upvotes · similarity 0.51
- Context-compact · hn · 2026-03-07 · 6 upvotes · similarity 0.50
- GibRAM an in-memory ephemeral GraphRAG runtime for retrieval · hn · 2026-01-18 · 60 upvotes · similarity 0.50
- LLM Memory Storage that scales, easily integrates, and is smart · hn · 2026-03-16 · 6 upvotes · similarity 0.49
- Fixing LLM memory degradation in long coding sessions · hn · 2025-11-27 · 5 upvotes · similarity 0.47
- AI memory with biological decay (52% recall) · hn · 2026-04-26 · 98 upvotes · similarity 0.47
- Llm.sql · hn · 2026-04-24 · 8 upvotes · similarity 0.47
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a model & infra tool for Fintech yet.