Weaviate Agents

Weaviate Agents are pre-built agentic services designed to simplify common tasks when working with Large Language Models (LLMs) and your Weaviate instance. They are experts in performing Weaviate-specific data operations, streamlining AI development and data engineering workflows.

Warning

Weaviate Agents are currently only available for users of Weaviate Cloud (WCD). They require a connection to a WCD instance and are not available for self-hosted Weaviate clusters at this time.

Available agents

The Weaviate Agents framework currently includes the following:

  • Query Agent: Designed to answer natural language questions by intelligently querying the data stored within your Weaviate database.

  • Transformation Agent: Enhances your data by manipulating it based on specific user instructions. This can involve tasks like summarizing, extracting information, or reformatting data stored in Weaviate.

  • Personalization Agent: Customizes outputs based on persona-specific information. This agent can learn user behavior and provide recommendations tailored to individual preferences.

Installation

To use the Weaviate Agents, you need to install the main weaviate-client package with the optional [agents] extra dependency. Run the following command:

pip install -U "weaviate-client[agents]"

This ensures you have both the core client and the necessary components for interacting with the agents.

Troubleshooting: Force pip to install the latest version

If you suspect you don’t have the latest agent features, or if instructed to use a specific version, you can try explicitly upgrading or installing the weaviate-agents package itself:

  • To upgrade to the latest available weaviate-agents version:

    pip install -U weaviate-agents

  • To install a specific weaviate-agents version:

    pip install -U weaviate-agents==<version_number>

    (Replace `<version_number>` with the desired version, e.g., `0.5.0`)

Official documentation

For the most comprehensive and up-to-date information, examples, and guides on Weaviate Agents, please refer to the main Weaviate documentation website:

Client API reference

Detailed documentation for the Python client classes and functions corresponding to these agents can be found below:

Support

  • Use our Forum for support or any other question.

  • Use our Slack Channel for discussions or any other question.

  • Use the weaviate tag on StackOverflow for questions.

  • For bugs or problems, submit a GitHub issue.