An enterprise-ready, local-first memory runtime with dated ingestion, hybrid recall, graph and entity operations, learning signals, cache and compression, and governance — per-workspace isolation, role-based access, GDPR controls, and a per-mode EU AI Act self-assessment. Optional network features are disclosed per capability.
Engineering capabilities documented across retrieval, storage, governance, and interfaces. Optional network features are disclosed separately per capability.
Five candidate producers feed fusion and reranking, with graph-derived score enhancement and inspectable diagnostics.
V3Public preprints evaluate information-geometric, topological, and stochastic-dynamics components. Current dense candidates use cosine, with research-derived terms applied only in documented later stages.
V3Per-workspace isolation, role-based access, GDPR access/erasure/portability, and a hash-chained audit trail. A per-mode EU AI Act self-assessment maps each mode (A/B local, C external). Engineering controls, not a certification.
V4SQLite with WAL, FTS, and profile isolation is the canonical path. CozoDB graph and LanceDB vector projections are staged, verified, and explicitly promoted by the Scale Engine.
Profile-selected MCP tools, structured CLI commands, and additive client assets support Codex, Claude Code, Cursor, Copilot, and Antigravity without overwriting a user's configuration.
Inspect Dashboard, Brain, Knowledge Graph, Memories, Health, Operations, Entity Explorer, Skill Evolution, Mesh Peers, Settings, and Optimize through the local service.
V3The answer check, which tells an agent whether the retrieved memories answer the question, is covered on Answer check. Memory kinds are covered under Workflows.
An inspectable stack from SQLite-backed storage through retrieval diagnostics, lifecycle controls, optional optimization, and local visualization.
SuperLocalMemory records local behavioral signals that influence ranking without a separate LLM call for the learning step. Three phases: baseline observation, rule-based heuristics, optional ML model activation. Inspect signals, thresholds, and rank changes; optional modes and integrations retain their own network paths.
Semantic, graph, and adaptive channels feed the ranking pipeline. Each channel is independently inspectable.
Dense vector candidates from a local embedding model. Finds memories conceptually related to a query even when keywords do not match exactly.
Entities — files, functions, concepts — are extracted and linked. Querying one entity can surface linked neighbours via graph traversal, not keyword matches alone.
Local behavioral signals — access patterns, explicit feedback, recency — adjust candidate rank without an additional LLM call for the learning step itself.
Recall emits local exposure telemetry. Exposure alone is not positive feedback and does not increase truth or trust scores stored in memory.
Optional reranking and channel diagnostics can record exposure. Explicit positive, negative, or corrective feedback remains separate from stored-memory confidence and trust. No model activation at this phase.
Heuristic boosts from learned patterns: recency weight, access frequency, and trust score. The ranking influence is deterministic and inspectable — no trained model is loaded at this phase.
When the learning path has sufficient valid signals and model activation is enabled, it can train on local behavioral state. Verify activation, model state, and rank impact in the installed release.
Each signal type has distinct semantics. Read the source before inferring intent from a rank change.
These features require hooks, background workers, or explicit configuration. Each needs independent verification for a complete deployment review.
Query enters, candidate producers run, re-ranking applies behavioral weights, ranked context exits. Every stage is traceable.
# Five candidate producers run in parallel stage candidates semantic (dense, local embedding) bm25 (full-text lexical) temporal (recency) hopfield (associative) spreading activation stage re-rank fusion score · entity-graph adjustment · access weight · trust score · decay ✓ ranked context returned · rank deltas inspectable in the dashboard # Per-channel score breakdown for any query $ slm trace "token expiry" names the channels that contributed to each result ✓ signal sources visible · no black-box ranking
Entity-graph information can adjust a post-fusion score but is not a separate candidate producer.
An exact cache path, content-aware compression controls, provider-prefix preparation, and release-gated semantic-cache safeguards. Cache reuse and compression results remain workload- and configuration-specific. Defaults: exact cache on, semantic cache off, safe compression on, aggressive compression off.
The primary Claude conversation turn cannot be cached without the proxy; the MCP tools and skills cache results you explicitly route through SLM. Optimization is designed to fail open: handled optimization errors pass through to the provider, though provider and transport failures remain possible.
The exact cache is the stable path. Semantic caching is opt-in because a similarity threshold can return the wrong response. The implementation contains threshold-learning machinery, but it is not a production claim until the frozen release proves calibration, invalidation, tenant isolation, and false-hit behavior on a representative corpus.
# Opt-in experimental path $ slm cache semantic on ↳ Enabled locally; this is not evidence of production precision # Inspect the cache $ slm cache status Entry count, DB size, TTLs, hit rate
Safe mode is lossless: whitespace and compact-JSON normalisation only. It never removes content and leaves code untouched, so reduction may be small or zero. Aggressive mode is opt-in and applies LLMLingua-2 to prose only, never code, numbers, structured data, instructions, or the current turn. A lossy result can be made reversible by storing its original in the CCR path and using the returned identifier before its TTL expires.
$ slm compress status ↳ Mode and per-channel state $ slm compress mode safe ↳ Lossless normalisation (aggressive is the opt-in alternative) $ slm compress prose on ↳ Prose compression, aggressive mode only
Alignment stabilises the prompt prefix to improve reuse of native provider prefix caching, and is on when compression is on. It helps only when the configured provider supports prompt caching and its cache conditions are met. Pricing, minimums, TTLs, and discounts belong to the provider's current contract; measure actual usage instead of publishing a fixed saving.
$ slm wrap claude ↳ Starts the proxy, sets ANTHROPIC_BASE_URL, launches Claude Code $ slm optimize status ↳ Show all settings (also: on, off) $ slm optimize savings --since 7 ↳ Token and cost report; compare against provider accounting ✓ No fixed hit-rate or cost multiplier without release-linked evidence
Read the full technical architecture or the published research behind the memory design, then follow the installation guide and compare cache correctness, token use, latency, and provider accounting on your own workload.