DealForge autonomously sources, scores, and writes investment memos on venture deals. Stop manually hunting.

1,180+ deals tracked  ·  22 AI investment memos  ·  Updated daily

← Back to leaderboard

Breathe

Show HN: Breathe-Memory – Associative memory injection for LLMs (not RAG)

40 AI Score
Show_hn other Added Mar 26, 2026

Details

Sector
other
Total Funding
$0
Last Round
$0

About

LLMs forget. The standard fix is RAG — retrieve chunks, stuff them in. It works until it doesn&#x27;t: irrelevant chunks waste tokens, summaries lose structure, and nothing actually models how memory works.<p>Breathe-memory takes a different approach: associative injection. Before each LLM call, it extracts anchors from the user&#x27;s message (entities, temporal references, emotional signals), traverses a concept graph via BFS, runs optional vector search, and injects only what&#x27;s relevant — typically in &lt;60ms.<p>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.<p>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.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;tkenaz&#x2F;breathe-memory" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;tkenaz&#x2F;breathe-memory</a><p>We&#x27;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.<p>We&#x27;ve also posted an article about memory injections in a more human-readable form, if you want to see the thinking under the hood: <a href="https:&#x2F;&#x2F;medium.com&#x2F;towards-artificial-intelligence&#x2F;beyond-rag-building-memory-injections-for-your-ai-assistants-ceedcea20419" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;towards-artificial-intelligence&#x2F;beyond-ra...</a>

AI Score Reasoning

Heuristic score based on available signals. Funding: $0, Source: show_hn.

Source

Show_hn — View original →