databricks-langchain

What it is and what it does

databricks-langchain bridges Databricks AI capabilities into LangChain applications, consolidating what were previously scattered across multiple packages. It provides LangChain-compatible wrappers for Databricks-hosted LLMs (like Llama and Mixtral via ChatDatabricks), vector search operations, embeddings generation, and experimental Genie integration for AI-powered analytics agents.

The package is designed for teams already using Databricks workspaces who want to build LangChain-based AI applications without managing separate authentication or API layers. It depends on langchain, databricks-sdk, and several Databricks-specific adapters (databricks-ai-bridge, databricks-mcp, databricks-openai, and others), making it a fairly heavy integration layer best suited for applications where Databricks is the primary AI infrastructure.

Use it for:

• Build LangChain agents that query Databricks-hosted LLMs like Llama 3.1 without direct SDK calls.
• Create RAG pipelines using DatabricksVectorSearch for semantic search over Databricks-managed vector stores.
• Generate embeddings for documents or queries using Databricks embedding models within LangChain workflows.
• Prototype Genie-powered analytics agents that answer business questions over Databricks data (preview feature).
• Migrate existing langchain-databricks or langchain-community Databricks code to the consolidated package.

Worth the install?

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

Integrates Databricks AI services—LLMs, vector search, embeddings, and Genie—into LangChain applications through a unified package.

Yes, if you are building LangChain applications on Databricks infrastructure. The package consolidates fragmented Databricks integrations and is actively maintained. The 11-dependency footprint is expected for an integration layer. No known vulnerabilities. Install only if you have a Databricks workspace and LangChain is your chosen framework; it adds no value as a standalone tool.

Install

databricks-langchain on PyPI

pip install databricks-langchain

uv add databricks-langchain

poetry add databricks-langchain

Installing databricks-langchain

Before you install

Low install friction; pure Python wheel. Active maintenance with recent release (65 days ago). Depends on 11 runtime packages including langchain, databricks-sdk, and several Databricks-specific bridges—a substantial dependency footprint typical of integration packages.

License in practice

Apache-2.0 permissive license allows commercial use, modification, and redistribution with minimal restrictions, suitable for most production scenarios.

Quickstart

pip install databricks-langchain

from databricks_langchain import ChatDatabricks

llm = ChatDatabricks(endpoint="databricks-meta-llama-3-1-70b-instruct")
response = llm.invoke("Hello")

Requires Python >=3.10 and valid Databricks workspace credentials to authenticate with Databricks endpoints.

Verify before relying

• Whether Genie Agent features are available outside the private preview or require special enablement.
• Performance characteristics and latency when calling Databricks endpoints through LangChain.
• Compatibility matrix with specific Databricks workspace versions or LangChain versions beyond the stated requirements.

Package facts

Evidence: databricks_langchain-0.20.0-py3-none-any.whl

Tags

More Artificial Intelligence packages