Pulls Latin dictionary data from English Wiktionary and morphological analysis from the Morpheus analyzer so your LLM can work with attested Latin words instead of hallucinating forms. Exposes three tools: search_latin takes any inflected form and returns definitions plus full paradigm tables, get_inflections fetches complete conjugation or declension data for a lemma, and lemmatize_latin analyzes a word like "amavit" into its grammatical components. Queries automatically resolve inflections to their base forms, macrons optional. Runs as a public streamable HTTP endpoint at latin.valksor.com/mcp with a parallel JSON API. Reach for this when you need an LLM to read, translate, or generate Latin with real grammatical grounding rather than plausible-sounding nonsense.