# SuperLocalMemory > Governed, local-first memory control plane for AI agents with MCP, an agent-native CLI, and auditable operations. SuperLocalMemory stores core memory state in a configurable local data root. Mode A does not require a cloud model provider for core memory operations. Optional providers, connectors, backup, proxies, model downloads, and dependency acquisition have separate network behavior. ## Current product contract - SuperLocalMemory is governed, local-first memory for AI agents: one long-term memory on the user's machine that learns from use, enforces who may read and erase what, coordinates many agents, and says "I don't have that" instead of guessing. - Answer check: Laya (on-device, Apple Silicon) or Jev (online, Windows, Linux and macOS, with consent) judges whether recalled memories answer the question; results are marked `abstained`: people see the closest memories marked related, while the agent recall tool returns nothing unless asked with `include_related`. - Web access (optional): ChatGPT, Claude on the web, Muse, Composio and MCP clients with remote OAuth connect through the Connected apps page; the memory database stays on the user's laptop. - Plugins: Claude Code, Codex, Antigravity, VS Code / Copilot, Hermes, Grok Bot; `slm connect` writes MCP config for the IDE you name. - Core memory is SQLite-backed; configuration, logs, queues, models, derived indexes, and optional backend state also live in the data root. - Recall uses five candidate producers followed by fusion, optional reranking, and graph-based score enhancement. - Dense candidate generation uses cosine similarity; Fisher-derived terms can inform later scoring. - Exact caching is available only when Optimize/proxy caching is enabled. - Safe compression preserves JSON and code; measured reduction varies by content and can be zero. Aggressive prose compression is experimental and lossy. - SQLite and sqlite-vec are canonical. Optional CozoDB and LanceDB projections remain explicit, parity-gated paths; they are not automatic replacements for the canonical store. - SLM governs each canonical write through admission and policy control, durable per-store obligations, and a hash-sealed completion manifest. A partial operation is surfaced as degraded rather than silently treated as complete. - Run `slm connect --list` for the 11 supported IDE targets. - SuperLocalMemory is AGPL-3.0-or-later. The score contract separates `relevance_score`, internal `ranking_score`, `memory_confidence`, and `trust_score`. Retrieval ordering is not answer probability. Until a held-out calibrator is frozen, canonical responses report `calibration_status: "uncalibrated"`, `calibration_id: null`, and `answer_confidence: null`. ## Research status SuperLocalMemory has four public arXiv preprints with code-linked experiments. They are not conference-accepted or journal-published. - https://arxiv.org/abs/2608.08253 - https://arxiv.org/abs/2603.02240 - https://arxiv.org/abs/2603.14588 - https://arxiv.org/abs/2604.04514 SuperLocalMemory 4.0: The Governed Memory Operating System for AI Agents is publicly available at https://arxiv.org/abs/2608.08253, with a companion Zenodo archive at https://zenodo.org/records/21853302 (DOI: https://doi.org/10.5281/zenodo.21853302). Published LoCoMo evidence from the architecture paper, with its original protocol scope: Mode A Raw scored 60.4% across 10 conversations / 1,276 questions with zero-LLM answer construction; Mode A Retrieval scored 74.8% across the same scope with local retrieval and GPT-4.1-mini answer synthesis; Mode C scored 87.7% on Conv-30 / 81 questions with cloud embeddings and GPT-4.1-mini answer generation and judge. These figures are not a fresh run of the current package. ## Links - Website: https://superlocalmemory.com - Launch overview: https://superlocalmemory.com/whats-new - Source: https://github.com/qualixar/superlocalmemory - Documentation: https://github.com/qualixar/superlocalmemory/wiki - PyPI: https://pypi.org/project/superlocalmemory/ - npm: https://www.npmjs.com/package/superlocalmemory