Skip to main content
The listLibraries tool returns all libraries indexed in your Grounded Docs server, along with detailed version information.

​
Tool Definition

MCP Tool Name: list_libraries Source: src/tools/ListLibrariesTool.ts Description: List all indexed libraries with version details including document counts, indexing status, and progress information.

​
Parameters

This tool takes no parameters.

​
Response Structure

​
libraries
LibraryInfo[]
Array of libraries with their version information
​
name
string
Library name
​
versions
array
Array of version objects
​
version
string
Version string (e.g., “18.2.0” or empty string for unversioned)
​
documentCount
number
Number of documents indexed for this version
​
uniqueUrlCount
number
Number of unique URLs indexed
​
indexedAt
string | null
ISO timestamp when indexing completed (null if in progress)
​
status
VersionStatus
Current status: "COMPLETED", "IN_PROGRESS", "FAILED", "QUEUED", or "CANCELLED"
​
progress
object
Progress information (only present for non-completed versions)
​
pages
number
Number of pages scraped so far
​
maxPages
number
Maximum pages configured for this scrape job
​
sourceUrl
string | null
Original documentation URL that was scraped

​
TypeScript Types

interface LibraryInfo {
  name: string;
  versions: Array<{
    version: string;
    documentCount: number;
    uniqueUrlCount: number;
    indexedAt: string | null;
    status: VersionStatus;
    progress?: { pages: number; maxPages: number };
    sourceUrl?: string | null;
  }>;
}

interface ListLibrariesResult {
  libraries: LibraryInfo[];
}

type VersionStatus = 
  | "COMPLETED" 
  | "IN_PROGRESS" 
  | "FAILED" 
  | "QUEUED" 
  | "CANCELLED";
See src/tools/ListLibrariesTool.ts:5-21 for the complete type definitions.

​
Example Request

{
  "name": "list_libraries",
  "arguments": {}
}

​
Example Response

​
With Multiple Libraries

{
  "libraries": [
    {
      "name": "react",
      "versions": [
        {
          "version": "18.2.0",
          "documentCount": 342,
          "uniqueUrlCount": 156,
          "indexedAt": "2024-03-15T10:30:00.000Z",
          "status": "COMPLETED",
          "sourceUrl": "https://react.dev/"
        },
        {
          "version": "17.0.2",
          "documentCount": 298,
          "uniqueUrlCount": 134,
          "indexedAt": "2024-02-20T08:15:00.000Z",
          "status": "COMPLETED",
          "sourceUrl": "https://legacy.reactjs.org/"
        }
      ]
    },
    {
      "name": "typescript",
      "versions": [
        {
          "version": "5.4.2",
          "documentCount": 124,
          "uniqueUrlCount": 89,
          "indexedAt": null,
          "status": "IN_PROGRESS",
          "progress": {
            "pages": 45,
            "maxPages": 100
          },
          "sourceUrl": "https://www.typescriptlang.org/docs/"
        }
      ]
    }
  ]
}

​
No Libraries Indexed

{
  "libraries": []
}

​
MCP Output

When called through the MCP interface (see src/mcp/mcpServer.ts:236), the response is formatted as: 
Indexed libraries:

- react
- typescript
- vue
If no libraries are indexed: 
No libraries indexed yet.

​
Usage from MCP Clients

​
Claude Desktop

List all indexed documentation libraries.

​
Direct MCP Call

import { Client } from '@modelcontextprotocol/sdk/client/index.js';

const result = await client.callTool({
  name: 'list_libraries',
  arguments: {}
});

​
Implementation Details

The tool retrieves library data from the document management service: 
async execute(_options?: Record<string, never>): Promise<ListLibrariesResult> {
  const rawLibraries = await this.docService.listLibraries();
  
  const libraries: LibraryInfo[] = rawLibraries.map(({ library, versions }) => ({
    name: library,
    versions: versions.map((v: VersionSummary) => ({
      version: v.ref.version,
      documentCount: v.counts.documents,
      uniqueUrlCount: v.counts.uniqueUrls,
      indexedAt: v.indexedAt,
      status: v.status,
      ...(v.progress ? { progress: v.progress } : undefined),
      sourceUrl: v.sourceUrl,
    })),
  }));
  
  return { libraries };
}
See src/tools/ListLibrariesTool.ts:33 for the complete implementation.

​
Version Status Flow

Libraries progress through these states during indexing: 
QUEUED → IN_PROGRESS → COMPLETED
                     ↘ FAILED
                     ↘ CANCELLED
  • QUEUED: Scraping job is waiting to start
  • IN_PROGRESS: Actively scraping and indexing pages
  • COMPLETED: All pages successfully indexed
  • FAILED: Scraping encountered an unrecoverable error
  • CANCELLED: Job was manually cancelled via cancel_job

​
Resource Alternative

You can also access libraries through the MCP resource endpoint: 
docs://libraries
This returns the same data but follows the MCP resource protocol. See src/mcp/mcpServer.ts:493 for implementation.

Build docs developers (and LLMs) love

Get started for freeTalk to us