AA
amalalexander.com_
back to articles
TOOLSMarch 18, 2026

I Turned 16 Months of Google Search Console Data Into a Vector DB

I had a simple question: what if I could talk directly to my search metrics? Here is how I loaded 16 months of queries and clicks into ChromaDB.

I Turned 16 Months of Google Search Console Data Into a Vector DB

Talking to Google Search Console Data

Storing raw search keywords in a spreadsheet is old school. By embedding 16 months of search performance data (queries, impressions, CTR, positions) into a vector database, we can perform semantic querying and automated clustering.

The Architecture

  • Data Source: Google Search Console API.
  • Embeddings Model: text-embedding-3-small (OpenAI).
  • Vector Database: ChromaDB (running locally).
  • Agent Interface: Streamlit + LangChain.

Now, queries like “What are my rising transactional intents that aren’t ranking on page 1?” return immediately with full context, bypassing complex Excel filters.

tags:
#vector db#search console#llm