Keyphrases
Extracts keyphrases from text using a BERT model, designed for integration with agentic applications and Model Context Protocol (MCP) clients.
About
This Keyphrases MCP Server provides a natural language interface for agentic applications to accurately extract keyphrases from text. It leverages the BERT machine learning model, integrating seamlessly with MCP clients to enhance AI-driven workflows. Addressing the limitations of autoregressive LLMs in comprehensive keyphrase contextualization, this server utilizes BERT's bidirectional processing to achieve higher relevance. It works directly with local files, exposing only the extracted keyphrases to the client while preserving document content integrity. The server combines BERT's extraction power with the generation capabilities of autoregressive LLMs, facilitating seamless text processing for various applications.
Key Features
- Utilizes KeyBERT framework for multi-step keyphrase extraction.
- Employs spaCy NLP for robust candidate generation and stop word discarding.
- Integrates with OpenAI, Claude Desktop, and OpenAI Agents SDK for flexible deployment.
- 0 GitHub stars
- Maximal Marginal Relevance (MMR) for diverse, non-redundant keyphrase selection.
- Semantic encoding with Sentence Transformers (e.g., paraphrase-multilingual-MiniLM-L12-v2).
Use Cases
- Generating tags or metadata for documents to improve organization and discoverability.
- Identifying emerging trends and sentiments from customer reviews, social media, or news articles.
- Providing features or inputs for downstream NLP tasks like text classification and clustering.
