Access Cohere language models for embeddings, classification, and generation.
## Cohere MCP Server: Enterprise NLP and RAG Platform The **Cohere MCP Server** integrates Cohere's enterprise-grade NLP capabilities into Google Antigravity. This platform specializes in embeddings, reranking, and retrieval-augmented generation (RAG), making it ideal for building production search and knowledge systems. ### Why Cohere MCP? Cohere excels at enterprise NLP applications: - **Enterprise Focus**: Built for production deployments - **Superior Embeddings**: Industry-leading text embeddings - **Reranking**: Improve search relevance significantly - **RAG Optimized**: Purpose-built for retrieval systems - **Multilingual**: Support for 100+ languages ### Key Features #### 1. Text Embeddings ```python import cohere co = cohere.Client("your-api-key") # Generate embeddings for semantic search response = co.embed( texts=["Machine learning is fascinating", "AI transforms industries"], model="embed-english-v3.0", input_type="search_document" ) embeddings = response.embeddings ``` #### 2. Reranking for Search ```python # Rerank search results for better relevance results = co.rerank( model="rerank-english-v3.0", query="How to optimize database queries?", documents=[ "SQL query optimization techniques", "Database indexing strategies", "NoSQL vs SQL databases" ], top_n=3 ) for result in results.results: print(f"{result.index}: {result.relevance_score}") ``` #### 3. RAG with Command-R ```python # Retrieval-augmented generation response = co.chat( model="command-r-plus", message="What are the key features of our product?", documents=[ {"title": "Product Guide", "snippet": "..."}, {"title": "FAQ", "snippet": "..."} ] ) # Response includes citations print(response.text) print(response.citations) ``` ### Configuration ```json { "mcpServers": { "cohere": { "command": "npx", "args": ["-y", "@anthropic/mcp-cohere"], "env": { "COHERE_API_KEY": "your-api-key" } } } } ``` ### Use Cases **Enterprise Search**: Build semantic search systems that understand user intent and return highly relevant results. **Knowledge Bases**: Create AI-powered documentation systems that answer questions with proper citations. **Content Classification**: Automatically categorize and tag content using Cohere's classification capabilities. The Cohere MCP Server brings enterprise-grade NLP to Antigravity for building production search and RAG applications.
{
"mcpServers": {
"cohere": {}
}
}