ZBS Index What actually exists in applied AI, with the source next to it

Memory and context

Everything here was classified as memory and context by keyword match against the maintainer's own description, so treat the grouping as a starting point rather than a verdict.

The list is ordered by the most recent commit, not by stars. A popular project that stopped in 2024 is not a better answer than a smaller one shipped last week.

Listed, not yet verified by us (60)

Published to the registry, but we have not yet checked its repository. Treat the entry as the maintainer’s claim only.

  • io.github.pintomatic/kernal-mcp — v0.2.1
    Graph-native memory for AI agents: a knowledge graph built from conversation, via MCP.
  • Cpsc — v0.1.0
    CPSC MCP — US consumer-product safety recalls (CPSC, free, no auth).
  • Diffbot — v0.1.0
    Diffbot MCP — Knowledge Graph company enrichment + web content extraction (diffbot.com)
  • Monarch Initiative — v0.1.0
    Monarch biomedical knowledge graph — diseases, phenotypes, genes, variants
  • Opentargets — v0.1.1
    Open Targets MCP — disease/target/drug knowledge graph.
  • SUBSTRATE — v1.0.1
    Persistent AI entity framework with causal memory, emotional state, and identity.
  • io.github.plur-ai/plur — v0.17.2
    Open, local-first engram memory for AI agents: read, correct, delete; shared over MCP.
  • GPU MCP Server — v0.1.0
    NVIDIA GPU metrics as MCP tools — utilization, memory, temperature, power. Supports MIG.
  • io.github.pmgarg/cgraphy — v0.1.1
    Code knowledge graph over MCP: token-budgeted context for AI coding agents.
  • Weasley DeepMind — v1.0.0
    Secure, fragment-based long-term memory for MCP-compatible AI agents.
  • io.github.pouyahasanamreji/continuum — v0.3.0
    Shared memory + orchestration for your coding agents. Local-first MCP, vector RAG.
  • MemoraEU — v0.1.6
    Zero-knowledge persistent memory layer for AI agents. Encrypted client-side, hosted in France.
  • MemoraEU — v0.1.2
    Zero-knowledge persistent memory for Claude — encrypted, semantic search, EU hosted.
  • io.github.pragneshbagary/memex — v0.1.9
    Persistent session memory for Claude Code. Local SQLite, no LLMs, no network.
  • io.github.pratikmehkarkar/vin-recall-mcp — v1.0.0
    Decodes US VINs and looks up open NHTSA safety recall campaigns for a vehicle.
  • io.github.praveen-ilangovan/pensieve — v0.1.0
    Manually-triggered, self-organising personal memory for AI agents; SQLite, no vector DB.
  • io.github.promptwheel-ai/logbook — v0.4.0
    Local Git-history memory for coding agents: reverts, suppressions, hotspots, and annotations.
  • io.github.Prophet-git/one-brain — v0.1.666
    Shared memory for a team in Claude Code: what one person records, everyone has.
  • io.github.pscale-commons/bsp-mcp — v0.4.0
    Persistent agent memory & identity on federated beaches — one bsp() function over pscale blocks.
  • Pseudolife Memory — v0.12.0
    Persistent memory for MCP-compatible agents: memory bank, fact cortex, dreams, graph.
  • io.github.purmemo-ai/purmemo — v12.9.5
    AI memory across Claude, ChatGPT, Gemini, and Cursor. Save, search, and recall context everywhere.
  • State Memory MCP — v0.9.31
    Deterministic, persistent graph server for tracking workflow state, decisions, and blockers.
  • Visual Memory MCP — v0.8.0
    Local visual UI cache for AI agents using perceptual hashing + CLIP to cut vision token use.
  • io.github.pyalwin/codemesh — v0.1.2
    Intelligent code knowledge graph for AI coding agents — 71% cheaper, 72% faster
  • CLS++ Memory — v7.19.0
    Cross-LLM persistent memory: store context once, recall it from any AI model.
  • io.github.RajdeepDas43/context-vault — v0.7.0
    Governed memory for AI agents (Paramind AI): conflict-checked writes, time-travel reads, ACL, audit.
  • NervaPack — v0.5.8
    Offline codebase knowledge graph: 91% token reduction vs naive RAG, plus cross-session agent memory.
  • NeuroStack — v0.12.0
    Token-efficient MCP memory for Markdown vaults. Tiered search, GraphRAG, AI memories.
  • io.github.rapozoantonio/memorykit — v1.4.2
    Cognitive memory for AI coding assistants — persistent memory across sessions
  • io.github.rarasAI/raras — v0.2.0
    Disease Twin — Brazil's rare-disease knowledge graph (10,468) over MCP. Source-grounded, PT-BR.
  • Rarefied Earth Groundwork MCP — v1.7.1
    Rarefied Earth Groundwork MCP: company memory for AI agents. 14-day full-access trial, no card.
  • Adaptive Memory Graph — v1.1.1
    Persistent memory for Claude via weighted, interconnected knowledge nodes.
  • io.github.rastogivaibhav/llm-kosh — v2.1.3
    Local-first AI memory cartridge with persistent MCP memory for Claude via SQLite FTS5.
  • io.github.ravi-labs/mindmap-mcp-server — v0.5.0
    Local-first memory & context-handoff across AI tools — capture in one, resume in another. No cloud.
  • state-trace — v0.3.1
    Bounded working memory for coding agents — typed nodes, current-vs-stale state, MCP-mountable.
  • io.github.RCOLKITT/vasperamemory — v1.0.0
    Persistent AI memory for coding assistants. 80+ tools for decisions, patterns, errors.
  • Tentra — Memory for AI Coding Agents — v1.3.2
    Memory for AI coding agents. Call graphs + refactor safety + persistent code context. 35 MCP tools.
  • io.github.RecallWorks/recall — v0.5.2
    Open-source MCP memory server. Persistent, searchable, tiered memory across sessions.
  • Remembra — AI Memory Layer — v0.12.1
    Persistent AI memory with entity resolution, temporal knowledge graph & 11 MCP tools.
  • Agentic Task System (ATS) — v0.9.0
    MCP server giving AI agents persistent task memory across TickTick, Notion, GitHub, Linear & Beads
  • MCP Memory Graph — v0.1.1
    Authority-weighted memory graph for AI agents with conflict detection and semantic search.
  • Yaps Local Memory — v1.0.0
    Search and safely manage your private local Yaps Markdown vault from MCP clients.
  • MeshMarket — v1.3.0
    The agent-to-agent capability exchange — rent memory, reasoning and safety, settled per call.
  • io.github.rigour-labs/rigour — v5.4.0
    Quality gates for AI agents. Lint, test, build checks with memory persistence.
  • io.github.riponcm/projectmem — v0.2.0
    Local-first memory + judgment layer for AI coding agents — warns before repeating failed fixes.
  • Recite MCP — v0.1.4
    Local MCP server for Recite receipt processing with ledger and memory tools.
  • io.github.robin-deruyck/mcp-ordali-memory — v1.0.1
    MCP server for Ordali memory: SQLite-backed project memory and governance traces.
  • robo-cortex — v0.4.1
    Git-aware memory store for AI coding agents; memories are anchored to git blob hashes.
  • Agent MCP Gateway — v0.2.5
    On-demand tool discovery for all your MCPs and per-subagent MCP controls to preserve context window.
  • io.github.rog0x/perf — v1.0.2
    Benchmark, memory, Big O analysis for AI agents
  • darwin-memo — v0.5.2
    Self-curating agent memory kept honest by survival-based selection over measured outcomes.
  • ZettelForge — v2.4.0
    Agentic memory for cyber threat intelligence. STIX graphs, actor aliasing, offline RAG, Sigma/YARA.
  • MCP OpenClaw Extensions — v3.3.0
    138-tool MCP server for AI agent firms: security, A2A, Hebbian memory, fleet mgmt
  • Memory OS AI — v3.0.1
    Adaptive memory for AI agents — FAISS search, chat extraction, cross-project linking
  • Ace Memory — v1.0.0
    Persistent memory for AI agents. Semantic search, memory graph, W3C DID identity.
  • Engram — v0.1.2
    Git-backed markdown memory for AI agents; search hides superseded and archived notes.
  • SNHP — free negotiation math + agent memory — v0.4.0
    Free game-theory negotiation advisor for agents, plus paid receipted sessions and agent memory.
  • io.github.sachitrafa/cognitive-ai-memory — v1.0.2
    Ebbinghaus-based persistent memory for Claude. Memories decay with time, strengthen on recall.
  • YourMemory — v1.4.18
    Persistent memory for AI agents. Memories decay with time and strengthen on recall.
  • io.github.SAIHM-Admin/saihm-mcp — v0.3.10
    Sovereign encrypted memory for AI agents. 8 MCP tools. GDPR erasure. Runnable offline demos.

Page 10 of 14

How this page is ordered

Entries are grouped by whether anyone is still working on them, using the date of the most recent push to the repository. They are not ordered by stars, because a star is a bookmark somebody left once and never took back.

Where we have not checked an entry yet, it says so rather than being mixed in with the verified ones.