How do I prepare json data for rag vectorstore?

Questions
1

I have a python code node that is producing some json output. See example below. What is the best way to process this output (if needed) to prepare for a vectorstore.

Secondly, what would be the best splitting method?I want to be able to use the cannonicalUrl and possibly the title as metadata.

Here is an example of the workflow, with some mock data:

Thanks.

2

To prepare your JSON data for a RAG vector store in n8n, start by extracting the `canonicalUrl` and `title` as metadata, and the `text` as the content. Use the **Vector Store** node with the **Insert Documents** operation to upload this data. For splitting, the **Recursive Character Text Splitter** is recommended for most cases, as it handles various formats like Markdown and HTML effectively. This setup ensures your data is ready for semantic searches. For more details, check the [docs.n8n.io]( RAG in n8n | n8n Docs ) guide on RAG.