Summary
Sentiment you can reproduce, because none of it is guessed.
The same word always encodes to the same code, and a text's valence comes back with the words that produced it — which makes the output something you can check rather than something you have to trust.
What it is
A Python MCP server over the Oyemi semantic lexicon. Every word maps to a fixed code in the form HHHH-LLLLL-P-A-V — superclass, lexeme, part of speech, abstractness, valence — and sentiment is computed from that lexicon rather than inferred, with no NLP libraries needed at runtime.
What you get
- A word encoded to a semantic code, with part of speech, abstractness and valence
- Many words encoded in one call
- A passage scored for valence, with the positive and negative words named
- A similarity score between two words, with a plain-language relationship label
- Synonyms and antonyms for a word
- The lexicon's name, version and word count
Requirements
Python, and pip install oyemi-mcp. No API key, no network calls, and no runtime NLP dependencies.
Setup effort
One command — pip install oyemi-mcp