@forgeframework/mcp-embeddings

MCP server for vector store and semantic search.

Part of the Forge Framework — a TypeScript framework for backend microservices and web applications.

Installation

npm install @forgeframework/mcp-embeddings

Usage

import { McpEmbeddingsServer } from '@forgeframework/mcp-embeddings';

const server = new McpEmbeddingsServer({
  vectorStore: {
    type: 'pgvector',
    connection: {
      host: 'localhost',
      port: 5432,
      database: 'vectors',
      username: 'embeddings',
      password: process.env.PGVECTOR_PASSWORD!,
    },
    defaultCollection: 'documents',
  },
  embeddingModel: {
    provider: 'openai',
    model: 'text-embedding-3-small',
    dimensions: 1536,
    apiKey: process.env.OPENAI_API_KEY!,
  },
  defaultTopK: 10,
  maxDocumentSize: 1048576,
  chunking: {
    strategy: 'sentence',
    maxChunkSize: 512,
    overlap: 50,
  },
});

await server.start({ type: 'stdio' });

Documentation

Full API reference and guides: https://carbonforge.io/framework/docs/mcp/mcp-embeddings

(Documentation site launching soon.)

License

MIT © Carbon Forge