MCP.so
ログイン

GemFireMCPServer

@jomartin-999

GemFireMCPServer について

概要はまだありません

基本情報

カテゴリ

その他

ライセンス

Apache-2.0 license

ランタイム

java

トランスポート

stdio

公開者

jomartin-999

設定

標準の設定はありません

このサーバーの README には解析可能な MCP 設定ブロックが含まれていません。インストール手順はリポジトリをご確認ください。

リポジトリ

ツール

ツールは検出されませんでした

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 GemFireMCPServer?

GemFireMCPServer is a Spring Boot–based MCP server that integrates VMware GemFire with Spring AI to enable fast, vector-based semantic search over financial documents. It is designed for developers using MCP clients like Claude Desktop to build Retrieval-Augmented Generation (RAG) workflows.

How to use GemFireMCPServer?

Clone the repository, configure GemFire artifact access credentials from Broadcom, set application.properties with GemFire cluster details and ONNX model paths, build with Gradle, start a GemFire cluster with the VectorDB extension, then register the server as an MCP tool in Claude Desktop. Three tools become available: add_financial_doc, list_available_financial_docs, and search_financial_docs.

Key features of GemFireMCPServer

  • MCP-compliant endpoints for document ingestion, listing, and querying
  • Local ONNX-based embeddings via Spring AI
  • Vector and metadata storage in VMware GemFire
  • File metadata stored in a dedicated GemFire region
  • Fast, in-memory semantic search

Use cases of GemFireMCPServer

  • Upload financial documents and automatically chunk, embed, and store them in GemFire
  • Browse metadata of all stored documents from an MCP client
  • Ask natural‑language questions about documents and receive RAG‑generated answers

FAQ from GemFireMCPServer

What runtime does GemFireMCPServer require?

Java 17+, Gradle (or Maven), VMware GemFire 10.x+, and the GemFire VectorDB extension are required. An MCP client like Claude Desktop is used to interact with the server.

How are embeddings generated?

The server uses a local ONNX‑exported version of the sentence-transformers/all-MiniLM-L6-v2 model. You must export the model with optimum-cli and place model.onnx and tokenizer.json on the classpath.

Where are documents and metadata stored?

All vector embeddings are stored in a GemFire vector index named financialDocuments. File metadata (name, size) is kept in a separate GemFire region defined by the gemfire.region.docsMetadata property.

How is the server invoked by the client?

The server uses the MCP STDIO transport, so it is launched as a subprocess by the MCP client (e.g., via java -jar in Claude Desktop’s configuration file). No HTTP server is involved.

What authentication is needed for GemFire dependencies?

Access to the GemFire artifact repository requires a Broadcom customer support username and a personal access token, which are configured in gradle.properties or Maven’s settings.xml.

コメント

「その他」の他のコンテンツ