OpenAI Agent Multiple Vectorstores

You can easily download and run the workflow directly in your KNIME installation. We recommend that you use the latest version of the KNIME Analytics Platform for optimal performance.

This workflow shows how to provide an OpenAI Functions Agent with multiple vector stores as tools.
When the agent is prompted, it can decide based on the input if it needs to retrieve information from the vector stores to respond to the prompt.

For now, the FAISS Vector Store Reader Node only supports local paths. Copy the YoutubeTranscripts and NodeDescriptions folders to a local path and point the FAISS Vector Store Reader to the folders to run the workflow.
Note that this agent requires the tool name to not contain white spaces.

In order to run the workflow you need an OpenAI API key. If you don't have one already, register with OpenAI and create a new API key under https://platform.openai.com/account/api-keys.

External resources

• Vector Stores LangChain Documentation
• Faiss LangChain Documentation
• OpenAI Functions Agent LangChain Documentation
• OpenAI API Keys