unitycatalog-langchain

What it is and what it does

This package bridges Unity Catalog functions and LangChain by wrapping UC functions as LangChain tools. It lets you register UC functions—whether hosted on a standalone UC server or in Databricks—and use them directly in LangChain agents and LLM tool-calling workflows. The toolkit handles function discovery, parameter binding, and execution, so agents can invoke UC functions as naturally as any other LangChain tool.

The package supports both generic Unity Catalog servers (via ApiClient configuration) and Databricks-managed UC (via DatabricksFunctionClient with workspace authentication). You define UC functions as Python callables or SQL, register them with the toolkit, and then pass them to LangChain agents. When an agent decides to call a UC function, the toolkit executes it and returns the result to the agent.

Use it for:

• Build agents that execute data transformations or analytics queries stored as UC functions without leaving the agent loop.
• Integrate Databricks SQL UDFs into LangChain workflows for real-time data lookups or computations during agent reasoning.
• Expose custom business logic (Python or SQL functions) as tools to LLMs, allowing agents to call them on demand.
• Create multi-step agentic workflows where UC functions serve as reusable, cataloged tools alongside other LangChain integrations.

Worth the install?

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

Exposes Unity Catalog functions as LangChain tools, allowing agents and LLMs to call UC functions directly within agentic workflows.

Yes, if you use LangChain agents and need to invoke Unity Catalog functions. The package has low install friction, permissive licensing, no known vulnerabilities, and is actively maintained. It fills a specific integration gap—making UC functions callable from agents—without adding significant complexity. Install it when you're already working with both LangChain and UC.

Install

unitycatalog-langchain on PyPI

pip install unitycatalog-langchain

uv add unitycatalog-langchain

poetry add unitycatalog-langchain

Installing unitycatalog-langchain

Before you install

Low friction install with three runtime dependencies (langchain, langchain-community, unitycatalog-ai). Actively maintained as of the latest release.

License in practice

Apache-2.0 permissive license allows commercial use, modification, and distribution with minimal restrictions.

Quickstart

pip install unitycatalog-langchain

from unitycatalog.ai.langchain.toolkit import UCFunctionToolkit

toolkit = UCFunctionToolkit(function_names=["catalog.schema.function_name"])
tools = toolkit.get_tools()

Requires Python 3.10 or later; also requires a running Unity Catalog server or Databricks workspace with appropriate authentication configured.

Verify before relying

• Whether the package supports LangGraph integration beyond what the description excerpt shows
• Performance characteristics when executing UC functions at scale through agents
• Compatibility guarantees with specific versions of langchain or langchain-community

Package facts

Evidence: unitycatalog_langchain-0.4.0-py3-none-any.whl

Tags

More Artificial Intelligence packages