Gujarati Lexicon MCP

What gujarati-lexicon-mcp does

The gujarati-lexicon-mcp MCP server supplies dictionary-backed Gujarati information to MCP-compatible AI assistants. It combines a hand-checked collection with data derived from Wiktionary, keeping the two sources identifiable in responses.

The define tool returns available meanings, part of speech, examples, and transliteration. synonyms gathers synonymous terms from the available sources and removes duplicates. idioms finds Gujarati idioms containing the requested word without generating entries that are not present in the data.

The package includes roughly 6,800 Wiktionary-derived words with English meanings. The curated data adds Gujarati meanings, idioms, and synonyms for a smaller set of entries. Lookups can also return “did you mean?” suggestions for likely typing errors.

How it works

A lookup first checks whether the input is an exact dictionary headword. When the input is an inflected form, the server can follow Wiktionary’s recorded relationship to a base word. If that information is unavailable, it removes selected Gujarati suffixes, including forms such as માં, નો, ની, and ે, and accepts the result only when it matches a real headword.

This approach supports examples such as ઘરમાં, ઘરે, આંખોમાં, and અમે. It can also follow some verb-form links, such as હોઈશ to હોવું. The server keeps curated and Wiktionary-derived results separate so an assistant can identify the source of a meaning.

The lexicon://word-of-the-day resource exposes a daily word from the hand-checked set. The explain_passage prompt uses the dictionary tools to explain a Gujarati passage one word at a time.

Setup and configuration

The gujarati-lexicon-mcp MCP server requires uv. For Claude Desktop, add gujarati-lexicon-mcp as the configured argument in the application’s developer configuration, then fully restart Claude Desktop before starting a new chat. On Windows, Claude Desktop may need the full path to uvx if it cannot locate that executable.

For local development, clone the repository, run uv sync, and use uv run pytest -v for tests. The project also documents uv run mcp dev src/gujarati_lexicon_mcp/server.py for opening the MCP Inspector.

Tools and capabilities

• define(word): returns dictionary meanings and related entry details.
• synonyms(word): returns deduplicated synonyms from the available sources.
• idioms(word): returns recorded idioms containing the word.
• lexicon://word-of-the-day: provides a daily word from curated data.
• explain_passage: explains Gujarati passages using the lookup tools.
• Inflection resolution, typo suggestions, transliterations, examples, and source attribution.

Limitations and notes

Gujarati verb coverage is incomplete: linked forms work in some cases, but phrases such as લે છે are not lemmatized. Wiktionary synonyms are combined across senses, so a word’s list may include a term associated with another meaning.

The curated collection is relatively small, and many entries contain English meanings rather than Gujarati explanations or idioms. Dictionary grounding applies to returned facts; it does not prevent the model from adding outside background knowledge, which may need to be labeled separately.

The project code is MIT-licensed. The included Wiktionary-derived data is licensed under CC BY-SA, so users should account for that distinction when redistributing the data.