fetch-embeddings

Retrieve embeddings for a file using its ID and a specific query, enabling vector search and context extraction via the Supavec MCP Server.

Instructions

Fetch embeddings for a file by ID and query

Input Schema

TableJSON Schema

Name	Required	Description	Default
`file_id`	Yes	ID of the file to get embeddings for
`query`	Yes	Query to search for in the file

Implementation Reference

src/tools/index.ts:69-106 (handler)

Executes the fetch-embeddings tool: extracts file_id and query from arguments, constructs API URL, calls makeSupavecRequest to Supavec embeddings endpoint, handles error or returns concatenated document contents as JSON text content.

if (request.params.name === "fetch-embeddings") {
  const file_id = request.params.arguments?.file_id as string;
  const query = request.params.arguments?.query as string;
  const embeddingsUrl = `${SUPAVEC_BASE_URL}/embeddings`;
  const embeddings = await makeSupavecRequest<Embeddings>(
    embeddingsUrl,
    {
      file_ids: [file_id],
      query: query,
    },
    apiKey
  );

  if ("error" in embeddings) {
    return {
      content: [
        {
          type: "text",
          text: `Failed to retrieve embeddings for ${file_id}: ${embeddings.error}`,
        },
      ],
    };
  }

  return {
    content: [
      {
        type: "text",
        mimeType: "application/json",
        text: JSON.stringify(
          embeddings.documents.map((d) => d.content).join("\n"),
          null,
          2
        ),
      },
    ],
  };
}

src/tools/index.ts:13-30 (registration)

Registers the fetch-embeddings tool in the tools array used for listTools response, including name, description, and inputSchema.

{
  name: "fetch-embeddings",
  description: "Fetch embeddings for a file by ID and query",
  inputSchema: {
    type: "object",
    properties: {
      file_id: {
        type: "string",
        description: "ID of the file to get embeddings for",
      },
      query: {
        type: "string",
        description: "Query to search for in the file",
      },
    },
    required: ["file_id", "query"],
  },
},

src/tools/index.ts:16-29 (schema)

Defines the JSON input schema for the fetch-embeddings tool, specifying required file_id and query parameters.

inputSchema: {
  type: "object",
  properties: {
    file_id: {
      type: "string",
      description: "ID of the file to get embeddings for",
    },
    query: {
      type: "string",
      description: "Query to search for in the file",
    },
  },
  required: ["file_id", "query"],
},

src/utils/api.ts:3-30 (helper)

Generic utility function to make authenticated POST requests to Supavec API endpoints, used by the fetch-embeddings handler to call the embeddings endpoint.

export async function makeSupavecRequest<T>(
  url: string,
  body: object,
  apiKey: string
): Promise<T | { error: string }> {
  try {
    const response = await fetch(url, {
      method: "POST",
      headers: {
        authorization: apiKey,
        "Content-Type": "application/json",
      },
      body: JSON.stringify(body),
    });
    if (!response.ok) {
      return {
        error: `Failed to fetch data: status ${response.status}`,
      };
    }

    const data = await response.json();
    return data as T;
  } catch (error) {
    return {
      error: `Failed to fetch data: ${error}`,
    };
  }
}

src/types/index.ts:1-5 (schema)
TypeScript type definition for the Embeddings response from the Supavec API, used in the handler.
```
export type Embeddings = {
  documents: {
    content: string;
  }[];
};
```

Supavec MCP Server