Skip to content
Agentic WorkFlow

Populate a Pinecone vector database from a website

Populate a Pinecone vector database from a website

Section titled “Populate a Pinecone vector database from a website”

Use Agentic WorkFlow to scrape a website, load the data into Pinecone, then query it using a chat workflow. This workflow uses the [HTTP node](/integrations/builtin/core-nodes/Agentic WorkFlow-nodes-base.httprequest/index.md) to get website data, extracts the relevant content using the [HTML node](/integrations/builtin/core-nodes/Agentic WorkFlow-nodes-base.html.md), then uses the Pinecone Vector Store node to send it to Pinecone.

[[ workflowDemo(“file:///advanced-ai/examples/populate_a_pinecone_vector_database_from_a_website.json”) ]]

Key features

Section titled “Key features”

This workflow uses:

  • [HTTP node](/integrations/builtin/core-nodes/Agentic WorkFlow-nodes-base.httprequest/index.md): fetches website data.
  • [HTML node](/integrations/builtin/core-nodes/Agentic WorkFlow-nodes-base.html.md): simplifies the data by extracting the main content from the page.
  • Pinecone Vector Store node and Embeddings OpenAI: transform the data into vectors and store it in Pinecone.
  • [Chat Trigger](/integrations/builtin/core-nodes/Agentic WorkFlow-nodes-langchain.chattrigger/index.md) and Question and Answer Chain to query the vector database.

Using the example

Section titled “Using the example”

—8<— “_snippets/examples-color-key.md”