広告掲載
MCP.so
ログイン
O

Openclaw Fcma MCP Server

@MohitDhawane

リポジトリ

基本情報

カテゴリ

メモリとナレッジ

トランスポート

stdio

作者

MohitDhawane

投稿者

MohitDhawane

Openclaw Fcma MCP Server について

6-tier AI memory system with 28 MCP tools — episodic memory, knowledge graphs, document storage, session ledger, auth & GDP 6-tier AI memory system with 28 MCP tools — episodic memory, knowledge graphs, document storage, session ledger, auth & GDPR complianceR compliance

Openclaw Fcma MCP Server の設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "openclaw-fcma": {
      "command": "python",
      "args": [
        "-m",
        "openclaw.mcp_server"
      ]
    }
  }
}

Openclaw Fcma MCP Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Openclaw Fcma?

Openclaw Fcma is a persistent memory layer for AI assistants, giving them structured recall across conversations. It uses a six-tier hierarchical architecture (scratchpad, journal, knowledge graph, document library, world state, and ledger) to store episodic memories, relationships, documents, live facts, and session summaries. Designed for developers and power users of Claude Desktop via the Model Context Protocol (MCP), it replaces flat retrieval-augmented generation (RAG) with surprise-driven learning and built-in user isolation.

How to use Openclaw Fcma?

Clone the repository, install dependencies with pip install -e ".[dev]", and add the server to Claude Desktop’s claude_desktop_config.json using python -m openclaw.mcp_server. After restarting Claude, use any of the 28 MCP tools—like memory_store, mind_add_node, library_upload, or user_register—to store and recall memories. The server can also run in HTTP mode with --transport http.

Key features of Openclaw Fcma

  • Six-tier hierarchical memory (L1–L6)
  • 28 MCP tools for episodic, graph, document, and key-value storage
  • Knowledge graph with neural decay and pruning
  • Surprise-driven memory updates with salience scoring
  • Built-in user management, JWT authentication, and rate limiting
  • GDPR-compliant data export and cascade deletion
  • Runs on FAISS vector store, SQLite, and Neo4j

Use cases of Openclaw Fcma

  • Give Claude persistent memory across conversations—remember user names, projects, and preferences.
  • Store and semantically search documents (PDF, text, code) with two-stage retrieval and re-ranking.
  • Build and query a knowledge graph of entities and relationships for complex reasoning.
  • Track live facts (key-value state) and session summaries for long-running projects.
  • Manage multi-user environments with per-user isolation, API keys, and audit logging.

よくある質問

1

What dependencies does Openclaw Fcma require?

Python 3.10+, FAISS, SQLAlchemy, Neo4j (for knowledge graph), and optional libraries for document ingestion. All dependencies are installed via `pip install -e ".[dev]"`.

2

Where is data stored?

Data is stored locally under a `data/` directory: SQLite database (`openclaw.db`), FAISS vector indexes (`journal/`, `library/`), Markdown session summaries (`ledger/`), and knowledge graph exports (`graph_exports/`). All paths are configurable via environment variables.

3

How does authentication work?

Openclaw Fcma uses JWT tokens with 24-hour expiry, auto-generated API keys (`fcma_xxxx…`), salted SHA-256 password hashing, and sliding-window rate limiting (default 60 requests per minute, configurable via `FCMA_RATE_LIMIT`).

4

What transports does the MCP server support?

The server supports stdio transport (default) and HTTP transport (`--transport http`). Use the MCP Inspector for debugging with `mcp dev openclaw/mcp_server.py`.

5

Can I export or delete my data for GDPR compliance?

Yes. The tools `user_data_export` and `user_data_delete` allow full export and cascade deletion of all user data, with an audit trail available via `audit_log`.

コメント

「メモリとナレッジ」の他のコンテンツ

UmmahAPI

Yousuf Kazmi

Islamic source data for assistants: Quran, hadith, tafsir, duas, prayer times and qibla. Answers come from real sources with citations instead of recalled text, so verses and hadith are not hallucinated. 36,000+ hadiths

Neither MCP

stonianua

Project context your AI can query through MCP — with source evidence.

Context7 MCP - Up-to-date Docs For Any Cursor Prompt

upstash

Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors

58.7K

server-noonien - shared memory MCP

sequico

Shared, serverless, conflict-free memory for AI agents: a multi-machine drop-in for the MCP memory knowledge graph.

Orcha

westonhancock

Orcha is a unified context layer for AI tools and agents. It stores and indexes and exposes organizational, team, or individual knowledge. Files, structured databases, and connected sources live in one workspace, and age

広告掲載