Tools: Pawz Engram biologically-inspired memory architecture for persistent AI agents

What is Engram? ## Why build this? ## The Engram loop (the part we care about) ## Architecture ## 1) Three tiers, three time scales ## 2) Hybrid retrieval (BM25 + vectors + graph) with fusion ## 3) Retrieval intelligence (a.k.a. don’t retrieve blindly) ## 4) Measured forgetting + safe rollback ## 5) Security by default ## What’s next ## Read the full whitepaper Most “agent memory” today is basically: Project Engram is our attempt to treat memory like a cognitive system instead of a dumping ground. It’s a three-tier architecture inspired by how humans handle information: Engram is implemented in OpenPawz, an open-source Tauri v2 desktop AI platform. Everything runs local-first. Read the full whitepaper: ENGRAM.md (Project Engram) Flat memory stores tend to fail the same way: Engram’s core bet is simple: Intelligent memory is not more memory — it’s better memory, injected at the right time and the right amount. Engram is built around a reinforcing loop: Tier 0: Sensory Buffer Tier 1: Working Memory Tier 2: Long-Term Memory Graph Engram fuses multiple signals rather than betting on one: Then it merges rankings with Reciprocal Rank Fusion (RRF) and can apply MMR for diversity. So weak results get corrected or rejected instead of injected as noise. Forgetting is first-class: That means storage stays lean without silently losing what matters. Engram encrypts sensitive fields before they hit disk: A few “high-leverage” additions we are actively working toward: If any of this resonates, the full architecture, modules, schema, and research mapping. And if you want to contribute, issues + PRs are welcome. Star the repo if you want to track progress. 🙏 Templates let you quickly answer FAQs or store snippets for re-use. Are you sure you want to hide this comment? It will become hidden in your post, but will still be visible via the comment's permalink. Hide child comments as well For further actions, you may consider blocking this person and/or reporting abuse - Sensory Buffer (Tier 0): short-lived, raw input for a single turn - Working Memory (Tier 1): what the agent is “currently aware of” under a strict token budget - Long-Term Memory Graph (Tier 2): persistent episodic + semantic + procedural memory with typed edges - everything competes equally (no prioritization) - nothing fades (stale facts stick around forever) - “facts”, “events”, and “how-to” get mixed into one blob pile - retrieval runs even when it shouldn’t (latency + context pollution) - security is often an afterthought - Gate: decide if memory retrieval is needed at all (skip trivial queries) - Retrieve: hybrid search (BM25 + vectors + graph signals) - Cap: budget-first context assembly (no overflow, no dilution) - Skill: store “how to do things” as procedural memory that compounds - Evaluate: track quality (NDCG, precision@k, latency) - Forget: measured decay + fusion + rollback if quality drops - FIFO ring buffer for this turn’s raw inputs (messages, tool outputs, recalled items) - drained into the prompt and then discarded - priority-evicted slots with a hard token budget - snapshots persist across agent switching - Episodic: what happened (sessions, outcomes, task results) - Semantic: what is true (subject–predicate–object triples) - Procedural: how to do things (step-by-step skills with success/failure tracking) - memories connect via typed edges (RelatedTo, Contradicts, Supports, FollowedBy, etc.) - BM25 for exactness and keyword reliability - Vector similarity when embeddings are available (optional) - Graph spreading activation to pull adjacent context - a Retrieval Gate: Skip / Retrieve / DeepRetrieve / Refuse / Defer - a Quality Gate (CRAG-style tiers): Correct / Ambiguous / Incorrect - decay follows a dual-layer model (fast-fade short memory vs slow-fade long memory) - near-duplicates are merged (“fusion”) - garbage collection is transactional: if retrieval quality drops beyond a threshold, Engram rolls back - automatic PII detection - AES-256-GCM field-level encryption - local-first storage design (no cloud vector DB dependency) - sensory buffer + working memory caches - graph store + typed edges - hybrid search + reranking - consolidation + fusion + decay + rollback - encryption + redaction defenses - observability (metrics + tracing) - proposition-level storage (atomic facts) - a stronger vector index backend (HNSW) - community / GraphRAG summaries for “global” queries - skill verification + compositional skills - evaluation harnesses (dilution testing + regression gates)