Local-first, audited memory MCP for Claude Code, OpenAI Codex, Gemini CLI, Cursor, Cline, Continue and OpenCode. Your agent gets a navigable cognitive map instead of a bloated context file. Zero cloud. SQLite local storage. Every mutation audited.

ArkHive — local-first governed memory for AI agents: tamper-evident SHA-256 ledger for identity, audit history, verification, and constraint-aware decisions. Apache-2.0.

Local-first, auditable memory for AI agents, with governed retrieval, delegated context delivery, and verifiable execution evidence. Works with OpenClaw, MCP, Pydantic AI, and LangChain/LangGraph. No external API required for local memory. Install with pip: pip install atmem

Agent platform with bitemporal memory, live observability, and audit-grade traceability — for AI that needs to act at the right moment

Agent memory you can audit. Every lesson cites its source; trace a bad memory and remove it with a receipt. Local-first MCP server for Claude Code, Cursor and Cline.

Desktop Extension for Novyx MCP — one-click install for Claude Desktop. Persistent memory for AI agents with rollback, audit trails, and semantic search.

Persistent memory + security layer for Claude Code (MCP plugin). Agents remember your project across sessions (~87% fewer context tokens), every tool call lands in a tamper-evident HMAC audit chain, and skills are AST-scanned + HMAC-verified before they run. 100% local: PostgreSQL + Ollama, no cloud, MIT.

Telarch for ChatGPT and Codex: a hosted MCP app at telarch.ai/mcp and a local stdio plugin for Harpe context budgets, Kibisis memory and Argus audit records.

MCP world-model memory server for coding agents; ships the etch-verify offline audit-chain verifier CLI

Forgetting engine for AI agents — memory TTL, decay, eviction, audit. The opposite of memory programming.

Native XMemo cloud-memory provider for OpenClaw agents: long-term memory, semantic recall, TODOs, snapshots, and audit tools.

Local-first memory governance for AI agents: shared, reviewable, auditable memory via SQLite and MCP.

Persistent Memory Kernel for AI Agents — crash recovery, shared memory, audit trail, real-time dashboard

The memory layer that can be corrected: provenance-first agent memory with a real corrections API and a one-command junk audit.

Auditable, local-first memory for AI agents in one SQLite file. MCP server for Claude Code, Cursor and Codex. npm: mnemosy-ai

Local-first, auditable memory layer for AI apps and coding agents — Codex, Claude Code, MCP, HTTP, TypeScript, and Python.

Open-source, self-hosted team memory for AI coding agents. Share, search, and audit project context through MCP.

Local-first persistent memory for Claude Code & Codex CLI - Rust CLI, hooks, MCP server, SQLite/SQLCipher, auditable recall for long-running coding work.

Local-first persistent memory and an auditable runtime for Claude, Cursor, Windsurf and any MCP client. Provenance, conflict handling, receipts and no cloud.

Governed local-first memory for coding agents — supersession, scoped recall, and audit on your machine. Spec v1.7.0, 143 validation behaviors.

Local-first MCP memory governance for coding agents — Codex, Claude Code, Cursor, Grok, and Trae. Shared rules, deduplication, token insights, audit, and rollback.

Agentic Event Store & Enterprise Audit Infrastructure: Building the immutable memory and governance backbone for multi-agent AI systems at production scale

Persistent cross-project memory for coding agents — an agent-authored markdown vault over MCP, auditable in Obsidian.

Local-first memory for DeepSeek Harness: per-turn BM25 recall you can audit — open-rate ledger, memory report, fail-closed write guard, 14-check doctor. Zero deps, no network, no model calls unless you opt in.

Caura (formerly MemClaw) — governed shared memory for AI agent fleets. Multi-agent, multi-tenant, MCP-native. Trust tiers, keystone policies, audit trails, knowledge graph, self-improving retrieval. Apache 2.0.

Local-first, self-hosted AI agent runtime and MCP bridge with sandboxed sessions, memory, credentials, audit/replay, and a local Console.

Standalone MCP server — cross-session memory, 10-gate VERITAS build pipeline, cryptographic audit ledger, Cortex approval gate. Two files, one dependency.

Persistent memory MCP server for AI agents (v0.18). 29 tools, self-growing graph without MCP sampling, poisoning quarantine: bitemporal facts, sqlx-migrate, BM25+MMR+OOD hybrid retrieval, graph metrics, conformal prediction, Hebbian/BCM, project scoping, tamper-evident audit chain (21 CFR Part 11-oriented), tree-sitter codegraph, git-native sync hooks. Rust + PostgreSQL + pgvector.

A local-first memory agent that remembers you. Everyday chats become auditable, Git-versioned Markdown. TUI + MCP.

Local-first verification layer for AI agents: evidence, provenance, policy gates, human approval, and auditable Trust Receipts before a claim, memory write, or action is trusted.

Persistent memory for AI coding agents. Local SQLite + UserPromptSubmit hook — matched beliefs in the prompt before the model sees it. Deterministic, auditable, no embeddings, no cloud.

Security memory with evidence attached. Human review, isolated client memories, and auditable recall in your own Cloudflare account. Built by Agent9.dev

Local memory server for AI agents over MCP: hybrid search, a fact graph with a duel rule, quarantine and a hash-chained audit log. Installs the prebuilt mgimind binary.

AI context that survives the chat window. Persistent repo memory, impact analysis, and instant style audits in one command.

The coordination and evidence layer for AI agent fleets: a local-first task board with atomic claims, live Mission Control, cost attribution, audit exports and shared memory for Claude Code, Codex and any MCP harness.

Local-first multilingual memory for Codex, Claude Code and MCP. Go + Qdrant + bge-m3, project isolation, verified full-backup rotation and preserved audit.

Auditable memory layer for AI agents: zero-LLM-call local ingest (~10ms/msg, air-gapped), matches Mem0 on accuracy at ~1000x lower ingest cost, bi-temporal belief-state, MCP server. Honest LoCoMo/LongMemEval benchmarks. Open source (Apache-2.0).

IJFW — It Just F*cking Works. Ferrox Labs' local-first infrastructure for AI coding agents: shared memory, smart routing, multi-AI cross-audits, disciplined workflow.

world-model-mcp is a signed audit memory server for AI coding agents, delivered as an MCP tool. Every event Ed25519-signed and Merkle-chained; offline-verifiable against pinned public keys. Hosted platform with hybrid post-quantum signing at etch.systems.

Let multibillion $ AI datacenters analyze the program memory for you. Create mods, trainers, security audits, game bots, accelerate RE, or do anything else with any program and game in a fraction of a time.

Agent memory you can audit. Every lesson cites its source, and a bad memory can be traced to where it came from and removed with a receipt. Local-first and open source, shared across Claude Code, Cursor and Cline over MCP.

Temporal Vault MCP: A local-first, security-hardened state management server for agentic workflows. Implements Content-Addressable Storage (CAS), Zstandard compression, and AES-256-GCM encryption to provide an auditable, "time-traveling" memory layer for LLM agents.

OCD (Omni Coder Dashboard) v5.6.0 — AI memory engine with real semantic embeddings, 25 MCP tools, session guard, trace-to-evidence audit bot, Antigravity intelligence, cost visibility, smart routing, and proactive IDE interception. 100% local, zero API keys.

HALO (Harness for Always-On Loop Oversight), from the survey Always-On Personal Intelligence: an evaluation harness and MCP plugin for always-on personal AI assistants and LLM agents. It checks whether Sensing, Memory, and Action are connected and whether self-evolving agents learn from outcomes, via paper and code audits and ablations.

AI-powered MCP server for JavaScript reverse engineering, security auditing, browser automation and dynamic analysis. Features CDP debugging, JS deobfuscation, network interception, memory scanning, Frida instrumentation, V8 inspector, WASM analysis, and cross-domain evidence correlation.

OCD (Omni Coder Dashboard) — An AI memory engine with real semantic embeddings, 21 MCP tools, cross-tool routing, trace auditing, and proactive IDE interception. 100% local, zero API keys required.

grāmatr — context engineering layer for AI coding agents. Every prompt gets a pre-computed intelligence packet: decision routing, capability audit, behavioral directives, memory pre-load, and ISC scaffolds. Continuity across sessions for Claude Code, Code

The policy layer for AI coding agents. aiZee governs what Cursor, Claude Code, Copilot, Windsurf, Devin & MCP agents may do — 139 enforcement gates, 29 personas, 245 skills, BMAD runtime, hybrid memory, tamper-evident audit. pip install aizee

Read-only Redis AI Assistant for Qoder IDE - MCP Server with 8 semantic tools for key inspection, search, pattern analysis, health check, memory analysis, slowlog analysis, cache advisor, and readonly queries. Features deep-defense security: command whitelist, key prefix filter, data masking, rate limiting, and audit logging.

Archivist MCP is a local-first memory server for coding agents. It persists decisions, incidents, rules, entities, and graph relationships in SQLite, then exposes them over MCP (stdio/HTTP), SSE team mode, and WebUI. Built for cross-session context continuity, conflict-safe updates, searchable recall, and auditable change history.

Apache Camel routes in Rust, and the reliable action layer for AI agents: routes as MCP tools with schema checks, approvals, run-once and audit. ~230 components, std-only, one native binary: 3 ms startup, up to 54x less memory, 1.7-8.7x Camel's throughput. A Cameleering project.

A text editor for LLMs, not humans. Take ed, the line editor that nobody has voluntarily used since about 1975, and rebuild it for an environment where the typing user is a language model. Short verbs, line addresses, no modes, no TUI. What an editor optimises for changes when the user is the model: round trips per task, tokens per command, an editing buffer with a long memory, and an undo tree that remembers the branches the agent abandoned, because that's often where the interesting work was. ## Features - **Fewer round trips.** Read-before-Edit is unnecessary; on conflict the response carries the new content so you reconcile in one call instead of pre-reading every time. - **Branching undo.** Walked-back work stays addressable instead of being thrown away when you pick a different path. - **Three-way merge.** Concurrent agents get a structured conflict response instead of a silent overwrite. - **Atomic batches.** Multi-file refactors run all-or-nothing instead of leaving half-applied state on failure. - **Cross-file moves and regex replace as primitives.** Operations the built-in tools can't express cleanly become single calls. - **Drift detection.** External edits to an open file are folded into the tree instead of being clobbered by the next write. - **Inline diagnostics.** Type errors and lint findings surface on save, not at the next build many edits later. - **Cross-session memory.** Per-file notes persist between sessions and surface inline on the next open. - **Audit log.** Every operation recorded with actor and timestamp, so two agents in one workspace can't argue about who moved the head. ## Multi-workspace (v0.4.0) One MCP server registration handles every project you mention. Tool calls route to the workspace owning the absolute path you pass; each project gets its own SQLite, its own undo tree, its own LSP daemon — lazily spawned on first touch. Desktop hosts (Claude Desktop, Codex Desktop, Cursor) install the server once globally and use it across projects without per-project config. ## Install `ae` must be installed locally; this MCP bundle registers `ae serve` as a stdio server. ```sh brew tap frane/tap && brew install agented # or curl -sSL https://raw.githubusercontent.com/frane/agented/master/install.sh | sh ``` Pure Go, single static binary, Apache 2.0. Source: https://github.com/frane/agented

Local-first cognitive map retrieval engine. SQLite + FTS5 + pluggable embeddings, with audited mutations and an offline consolidator.

—

Audit trail database for AI agents. Tamper-evident, checksum-backed decision logs with session replay and time-travel debugging to support EU AI Act Article 12 record-keeping. Drift detection, MCP, Python & TypeScript SDKs. Self-host or cloud.

内网 AI 知识库引擎（忆桥）· 企业文档问答 / 制度与合同检索 / 私有化部署 · 每个答案带逐句出处与可回放 traceId，版本链 + 审计 + 部门隔离 · Evidence-cited enterprise knowledge base & long-term memory MCP server for internal documents, local-first on SQLite + Ollama.

🧠 给 AI 一个可信的第二大脑：本地 SQLite · 先审核再入库 · 不可篡改审计 · 记忆账本。🕷 蜘蛛侠皮肤 · 🔍 自建集 R@5 99.1% / LongMemEval-S 97.2% · 🧩 一个库接入 DSH / Claude Code / Codex / Cursor / 任意 MCP · ⭐ MIT 开源 · 🚀 零依赖开箱即用

Provenance-aware memory format for AI agents: every claim carries its source, age and boundary.

Persistent memory + workflow orchestration as an MCP server. 78 canonical tools across 21 namespaces, SQLite-backed, with an LLM-as-judge drift gate and a vibe-flow loop (spec to artifact to drift to publish).

Lossless, compressed, verifiable memory for AI agents.