h2ogpte

What it is and what it does

h2ogpte is a Python client for H2OGPTe, an enterprise retrieval-augmented generation (RAG) platform. It provides programmatic access to create collections, upload and ingest documents, and run semantic queries against those documents using large language models. The client supports both global and collection-specific API keys for different levels of access control, and it can route queries to different LLMs, apply cost controls, enforce structured outputs (JSON or classification), and enable vision-capable modes when needed.

The package is designed for developers building applications that need to combine private or proprietary documents with LLM capabilities—for example, contract analysis, knowledge base Q&A, or document summarization. It handles the async communication layer via aiohttp and websockets, document parsing via beautifulsoup4 and lxml, and structured response formatting via pydantic. The client also provides an OpenAI-compatible API endpoint, allowing it to work with standard OpenAI client libraries.

Use it for:

• Upload and query a collection of contracts or legal documents to extract specific terms or answer compliance questions.
• Build a chatbot that answers questions about internal documentation or knowledge bases by ingesting and searching those documents.
• Summarize multiple documents in a collection automatically using the document processing API.
• Route queries to cost-optimized LLMs based on performance requirements and budget constraints.
• Integrate LLM-powered Q&A into existing applications via the OpenAI-compatible REST API.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Python client for querying and managing documents in H2OGPTe, an enterprise RAG platform that combines document ingestion, semantic search, and LLM-powered question-answering.

Yes, if you are building on H2OGPTe. The client is actively maintained, has low install friction, supports Python 3.8 through 3.12, and provides a complete interface to the platform's core features. The unclear license status and lack of public visibility into dependency availability warrant checking H2O's documentation before committing to a production deployment. No known security vulnerabilities.

Install

h2ogpte on PyPI

pip install h2ogpte

uv add h2ogpte

poetry add h2ogpte

Installing h2ogpte

Before you install

Low friction installation as a pure Python wheel. Active maintenance with a release within the last day. Supports Python 3.8 through 3.12 and PyPy. Depends on 20 runtime packages including aiohttp, pydantic, pandas, and websockets—a moderate but standard dependency footprint for an async-capable client library.

Quickstart

pip install h2ogpte

from h2ogpte import H2OGPTE

client = H2OGPTE(
    address='https://h2ogpte.genai.h2o.ai',
    api_key='sk-XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX',
)

collection_id = client.create_collection(
    name='Contracts',
    description='Paper clip supply contracts',
)

chat_session_id = client.create_chat_session(collection_id)

with client.connect(chat_session_id) as session:
    reply = session.query(
        'How many paper clips were shipped to Scranton?',
        timeout=60,
    )
    print(reply.content)

Requires a valid API key and network access to an H2OGPTe instance at the specified address.

Verify before relying

• Whether the package's license is proprietary, open-source, or dual-licensed—the fact sheet does not specify.
• Whether h2o_authn and other H2O-specific dependencies are publicly available or require separate registration.
• Performance characteristics and typical latency for document ingestion and query operations.

Package facts

Evidence: h2ogpte-1.7.4-py3-none-any.whl

Keywords: information-retrieval, LLM, large-language-models, question-answering, search, semantic-search, analytical-search, lexical-search, document-search, natural-language-querying

Tags

More Artificial Intelligence packages