Breathe-Memory – Associative memory injection for LLMs (not RAG)
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,…
What it does
In the maker’s words, at launch
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-memory We'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...
Does the same job
all alternatives →
- ZLZep – Long-Term Memory Store for LLM Apps2023 · ▲7
Zep is a long-term memory store designed for conversational AI applications built using modern LLMs. It handles the storage, summarization, embedding, indexing, and enrichment of chat histories, and offers developers a simple, low-latency API to this data. Chat history storage is an infrastructure challenge all developers and enterprises face as they look to move from prototypes to deploying conversational AI applications that provide rich and intimate experiences to users. Key features include long-term memory persistence, auto-summarization, vector search, auto-token counting, and Python…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
