langchain-nvidia-ai-endpoints

What it is and what it does

This package bridges LangChain and NVIDIA's AI infrastructure, letting you use NVIDIA Foundation Models—particularly Nemotron and other open models—as chat and embedding providers. It works by connecting to either live endpoints on the NVIDIA API Catalog (cloud-hosted on DGX infrastructure) or to self-hosted NIM microservices running in containers on your own infrastructure.

You instantiate a ChatNVIDIA object with a model name, then use standard LangChain interfaces: invoke, stream, batch, and their async variants all work natively. The package handles authentication via an NVIDIA_API_KEY environment variable and exposes available_models to discover which models your credentials can access. It integrates with LangChain's prompt templates and output parsers, so you can build chains and agents using NVIDIA models as the LLM backbone.

Use it for:

• Build agentic AI workflows using Nemotron's reasoning and tool-calling capabilities within LangChain
• Generate code using specialized models like meta/codellama-70b or google/codegemma-7b through LangChain chains
• Process multimodal inputs (text and images) with models like nvidia/neva-22b for reasoning tasks
• Stream real-time responses from NVIDIA models in LangChain applications without blocking
• Run inference on-premises using self-hosted NIM containers for full IP and customization control

Worth the install?

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

Integrates NVIDIA AI Foundation Models and chat endpoints into LangChain applications, providing access to models like Nemotron through the NVIDIA API Catalog or self-hosted NIM containers.

Yes. The package has low install friction, active maintenance, no known vulnerabilities, permissive MIT licensing, and integrates seamlessly into LangChain workflows. It is worth installing if you need access to NVIDIA Foundation Models (especially Nemotron) or want the option to run models on-premises via NIM.

Install

langchain-nvidia-ai-endpoints on PyPI

pip install langchain-nvidia-ai-endpoints

uv add langchain-nvidia-ai-endpoints

poetry add langchain-nvidia-ai-endpoints

Installing langchain-nvidia-ai-endpoints

Before you install

Low friction installation with three runtime dependencies (aiohttp, langchain-core, requests). Package is actively maintained with recent commits and no known vulnerabilities.

License in practice

MIT license permits free use, modification, and distribution with minimal restrictions, making it suitable for both open-source and commercial projects.

Quickstart

pip install langchain-nvidia-ai-endpoints

from langchain_nvidia_ai_endpoints import ChatNVIDIA

llm = ChatNVIDIA(model="nvidia/nemotron-3-super-120b-a12b")
result = llm.invoke("Write a ballad about LangChain.")
print(result.content)

Requires NVIDIA_API_KEY environment variable set to a valid key from https://build.nvidia.com/ (format: nvapi-*), or a running NIM container endpoint if using self-hosted models.

Verify before relying

• Whether the package supports all NVIDIA Foundation Models listed in the API Catalog or only a subset
• Performance characteristics and latency expectations when using the NVIDIA API Catalog versus self-hosted NIM
• Cost implications of using the NVIDIA API Catalog endpoints

Package facts

Evidence: langchain_nvidia_ai_endpoints-1.4.3-py3-none-any.whl

Tags

More Artificial Intelligence packages