Cohere API Command R Command R Plus Embed Integration

Cohere API Integration: Command R, Command R+, Embed

Cohere specializes in enterprise NLP: their embeddings (embed-multilingual-v3) lead the MTEB benchmark for multilingual search. Command R+ is optimized for RAG tasks with built-in RAG mode that returns source citations. Useful for enterprise search requiring verifiable answers.

Basic Integration

import cohere

co = cohere.Client("COHERE_API_KEY")

# Chat (Command R+)
response = co.chat(
    model="command-r-plus",
    message="Explain the principles of transformers",
    temperature=0.1,
)
print(response.text)

# Async
import cohere.asyncio as async_cohere

async_co = async_cohere.AsyncClient("COHERE_API_KEY")
response = await async_co.chat(model="command-r-plus", message="Request")

RAG Mode with Built-in Citations

documents = [
    {"id": "doc_1", "title": "Security Policy", "text": "...text..."},
    {"id": "doc_2", "title": "Access Rules", "text": "...text..."},
]

# Cohere automatically finds relevant documents and cites them
response = co.chat(
    model="command-r-plus",
    message="How to get access to corporate systems?",
    documents=documents,
    # Response contains citations with source references
)

print(response.text)
for citation in response.citations:
    print(f"Citation: {citation.text}, sources: {citation.document_ids}")

Embeddings (Best in Class for Search)

# Multilingual embeddings — one of best options for RU/EN/UA
response = co.embed(
    texts=["Search documents", "Document search", "Пошук документів"],
    model="embed-multilingual-v3.0",
    input_type="search_query",  # "search_query" or "search_document"
)
embeddings = response.embeddings

# For indexing documents
doc_embeddings = co.embed(
    texts=["Document text 1", "Document text 2"],
    model="embed-multilingual-v3.0",
    input_type="search_document",
)

Rerank — Search Results Re-ranking

# Cohere Rerank — powerful tool for improving RAG accuracy
docs = [
    "Python is an interpreted programming language",
    "Anaconda is a Python distribution for data science",
    "Python snakes are widespread in tropical regions",
    "Django is a Python web framework",
]

results = co.rerank(
    model="rerank-multilingual-v3.0",
    query="Python for machine learning",
    documents=docs,
    top_n=3,
)

for result in results.results:
    print(f"Score: {result.relevance_score:.3f} | {docs[result.index]}")

Cost of Cohere (2025)

Timeline

• Basic integration of chat: 0.5–1 day
• RAG with citations: 2–3 days
• Rerank pipeline: 1–2 days