Kgrag mcp server (github#248)

* feat: aggiungere configurazione del server Knowledge Graph MCP

* feat: aggiornare la configurazione del server KGraph MCP e aggiungere strumenti per query, ingestione ed estrazione

Author: gzileni

servers/kgrag-mcp-server/server.yaml

@@ -0,0 +1,130 @@
+name: kgrag-mcp-server
+image: mcp/kgrag_mcp_server
+type: server
+meta:
+  category: productivity
+  tags:
+    - knowledge-graph
+    - search
+    - productivity
+about:
+  title: Knowledge Graph MCP Server
+  description: |-
+    Manage, ingest, and query structured/unstructured data.
+    Designed to integrate with Neo4j, AWS S3, Redis, Qdrant, and LLMs.
+  icon: https://avatars.githubusercontent.com/u/6775950?v=4
+
+source:
+  project: https://github.com/gzileni/kgrag_mcp_server
+  branch: main
+
+config:
+  description: Configure the connection to the Knowledge Graph MCP Server
+  env:
+    - name: LLM_MODEL_TYPE
+      example: ollama
+      value: '{{kgrag-mcp-server.llm_model_type}}'
+    - name: LLM_MODEL_NAME
+      example: llama3.2
+      value: '{{kgrag-mcp-server.llm_model_name}}'
+    - name: API_KEY
+      example: sk-...
+      value: '{{kgrag-mcp-server.api_key}}'
+    - name: LLM_URL
+      example: http://localhost:11434
+      value: '{{kgrag-mcp-server.llm_url}}'
+    - name: TEMPERATURE
+      example: "0.5"
+      value: '{{kgrag-mcp-server.temperature}}'
+
+    # Vectors & Collections
+    - name: VECTORDB_SENTENCE_MODEL
+      example: BAAI/bge-small-en-v1.5
+      value: '{{kgrag-mcp-server.vectordb_sentence_model}}'
+    - name: COLLECTION_NAME
+      example: kgrag_data
+      value: '{{kgrag-mcp-server.collection_name}}'
+
+    # Model embedding
+    - name: MODEL_EMBEDDING
+      example: nomic-embed-text
+      value: '{{kgrag-mcp-server.model_embedding}}'
+    - name: LLM_EMBEDDING_URL
+      example: http://localhost:11434
+      value: '{{kgrag-mcp-server.llm_embedding_url}}'
+
+    # AWS
+    - name: AWS_ACCESS_KEY_ID
+      example: AK...
+      value: '{{kgrag-mcp-server.aws_access_key_id}}'
+    - name: AWS_SECRET_ACCESS_KEY
+      example: cG....
+      value: '{{kgrag-mcp-server.aws_secret_access_key}}'
+    - name: AWS_REGION
+      example: eu-central-1
+      value: '{{kgrag-mcp-server.aws_region}}'
+    - name: AWS_BUCKET_NAME
+      example: my_bucket
+      value: '{{kgrag-mcp-server.aws_bucket_name}}'
+
+  parameters:
+  type: object
+  properties:
+    llm_model_type:
+      type: string
+      default: ollama
+      description: LLM provider (e.g., ollama, openai)
+    llm_model_name:
+      type: string
+      default: llama3.2
+      description: Name of the LLM model to use
+    api_key:
+      type: string
+      description: API key for the LLM provider (not required for local Ollama)
+    llm_url:
+      type: string
+      default: http://localhost:11434
+      description: Base URL of the LLM provider or gateway (e.g., Ollama/OpenAI proxy)
+    temperature:
+      type: number
+      default: 0.5
+      description: Sampling temperature (range 0–2)
+
+    # Vectors & Collections
+    vectordb_sentence_model:
+      type: string
+      default: BAAI/bge-small-en-v1.5
+      description: Sentence embedding model for the vector database
+    collection_name:
+      type: string
+      default: kgrag_data
+      description: Name of the collection/index in the vector database
+
+    # Model embedding
+    model_embedding:
+      type: string
+      default: nomic-embed-text
+      description: Model used for generating embeddings
+    llm_embedding_url:
+      type: string
+      default: http://localhost:11434
+      description: Endpoint for generating embeddings
+
+    # AWS (optional)
+    aws_access_key_id:
+      type: string
+      description: AWS Access Key ID
+    aws_secret_access_key:
+      type: string
+      description: AWS Secret Access Key
+    aws_region:
+      type: string
+      default: eu-central-1
+      description: AWS region
+    aws_bucket_name:
+      type: string
+      description: AWS S3 bucket name
+  required:
+    - llm_model_type
+    - llm_model_name
+

servers/kgrag-mcp-server/tools.json

@@ -0,0 +1,35 @@
+[
+  {
+    "name": "query",
+    "description": "Query the KGraph system with a specific query string.",
+    "arguments": [
+      {
+        "name": "prompt",
+        "type": "string",
+        "desc": "The query string to search for."
+      }
+    ]
+  },
+  {
+    "name": "ingestion",
+    "description": "Ingest a path of file into the KGraph system.",
+    "arguments": [
+      {
+        "name": "path",
+        "type": "string",
+        "desc": "The file path to ingest into the KGraph system."
+      }
+    ]
+  },
+  {
+    "name": "extract",
+    "description": "Extract graph data from a document using the KGraph system.",
+    "arguments": [
+      {
+        "name": "text",
+        "type": "string",
+        "desc": "The nodes, entities and relations to extract information from."
+      }
+    ]
+  }
+]