SuperLocalMemory 4.0: The Governed Memory Operating System for AI Agents

SuperLocalMemory 4.0 is an open, local-first memory operating system for AI agents. It brings retrieval, learning, governance, time awareness, and operational reliability into one control plane that organizations can run on infrastructure they control. The system combines five candidate-producing retrieval channels—dense semantic, BM25 lexical, temporal, Hopfield-associative, and spreading-activation retrieval—fused through reciprocal-rank fusion. It adds a governed learning and behaviour layer, bi-temporal recall, multi-scope memory for personal, shared, and global contexts, role-based access control, GDPR-oriented export and verified erasure, a hash-chained audit trail, and an EU AI Act deployment-context checklist. V4 introduces a reliability spine for the primary write path: generation-fenced admission, a policy registry, verifiable memory transactions with per-projection apply, verify, compensate, and erase owners, and hash-checkable completion manifests. The implementation is available through CLI, MCP, HTTP daemon, dashboard, editor integration, and framework adapters; it supports fully local, local-with-on-device-model, and provider-assisted operating modes. The released reliability evidence bundle reports 2,200 of 2,200 deterministic repetitions upholding their scoped component properties across eleven fault-injection and mechanism scenarios. The governed write envelope measured 3.522 ms at p50 and 5.297 ms at p99, versus 1.835 ms and 2.569 ms for the ungoverned baseline, for an in-process control-plane overhead of 1.687 ms at p50 and 2.728 ms at p99. These are scoped component and mechanism measurements, not an end-to-end multi-process or external retrieval-accuracy benchmark. This work consolidates three prior SuperLocalMemory preprints: privacy-preserving multi-agent memory and Bayesian trust defence; information-geometric retrieval foundations; and the V3.3 Living Brain lifecycle and multi-channel retrieval work. The exact v4.0.0 software release is available at https://github.com/qualixar/superlocalmemory/releases/tag/v4.0.0. Lead author research profile: https://arxiv.org/a/bhardwaj_v_1.html