NLI Search

Name: NLI Search
Author: mula2812

Community

MCP server enabling natural language search across millions of cultural, historical, and literary assets held by the National Library of Israel (NLI). Converts conversational queries in Hebrew or English into structured API calls against the NLI digital archive, returning results with thumbnails, IIIF manifests, and streaming content. The collection includes ancient manuscripts, government records, photographs, maps, music recordings, oral histories, and scholarly publications spanning Jewish heritage and Israeli culture.

3 Stars6 tools0 installs34 viewsLibrary & Archives

Agent Configuration

Clone from GitHub to install

git clone https://github.com/mula2812/NLI_AI_Search
cd NLI_AI_Search
pip install -r requirements.txt

.mcp.json

{
  "mcpServers": {
    "nli-search": {
      "command": "python3",
      "args": [
        "nli_mcp.py"
      ],
      "cwd": "/path/to/NLI_AI_Search"
    }
  }
}

Features

Natural language search across the National Library archives
Bilingual support (Hebrew/English) for queries and responses
Automatic thumbnail extraction with IIIF fallback
Support for historical documents, literature, photographs, audio, and video
Streaming batch processing for large result sets
Media access: MP4, HLS, and audio streams

Tools(6)

process_natural_query
Process natural language queries and convert them to structured search parameters for the NLI API
generate_response
Generate a user-facing response from search results with image extraction and formatting
stream_batches
Process large result sets in streaming batches for comprehensive analysis
get_image
Retrieve an image by IIIF identifier
get_manifest
Retrieve the IIIF manifest for a given record ID
get_stream
Retrieve media streams (MP4, HLS, audio) for a library item

Data Sources(2)

National Library of Israel OpenLibrary API
Primary search API for the National Library of Israel digital archives
IIIF Image API
International Image Interoperability Framework endpoints for images and manifests

Try These Queries

Search for books by Bialik in the National Library

Find historical photographs of Jerusalem

Show ancient manuscripts about Kabbalah

Search for Leah Goldberg children poems

Find audio recordings of Israeli Mizrahi music

Setup Requirements

Environment Variables

NLI_API_KEY (optional, default visitor key provided)

Dependencies

Python 3.10+mcphttpxpydantic

Notes

A default visitor API key is included. Installation scripts provided for Windows and Linux. Installation path must contain only English characters.

Known Limitations

Requires Claude Desktop version 2.0 or higher
Installation path must contain only English characters
Commercial use requires a separate licensing agreement
Not officially affiliated with the National Library of Israel

Author

mula2812

Language

Python

License

MIT

Stars

Transport

Standard I/O (stdio)

libraryarchivebooksmanuscriptsIIIFcultural heritage

View on GitHub

Agent Configuration

Clone from GitHub to install

git clone https://github.com/mula2812/NLI_AI_Search
cd NLI_AI_Search
pip install -r requirements.txt

.mcp.json

{
  "mcpServers": {
    "nli-search": {
      "command": "python3",
      "args": [
        "nli_mcp.py"
      ],
      "cwd": "/path/to/NLI_AI_Search"
    }
  }
}

Features

Natural language search across the National Library archives

Bilingual support (Hebrew/English) for queries and responses

Automatic thumbnail extraction with IIIF fallback

Support for historical documents, literature, photographs, audio, and video

Streaming batch processing for large result sets

Media access: MP4, HLS, and audio streams

Tools(6)

process_natural_query

Process natural language queries and convert them to structured search parameters for the NLI API

generate_response

Generate a user-facing response from search results with image extraction and formatting

stream_batches

Process large result sets in streaming batches for comprehensive analysis

get_image

Retrieve an image by IIIF identifier

get_manifest

Retrieve the IIIF manifest for a given record ID

get_stream

Retrieve media streams (MP4, HLS, audio) for a library item