google-image-search

You can automate image search by using Google Custom Search API to fetch deterministic results and an LLM to score and select the best visuals for note enrichment. This approach surfaces high-quality images without manual browsing.

LLM-based scoring evaluates images fetched via Google Custom Search API to rank and select the most suitable visual per topic. This mechanism ensures deterministic results are filtered for relevance and quality.

Yes, you need a Google Custom Search API key and CX to perform the image search. You also need access to an LLM via OpenRouter with a local llm executable for the selection process.

Yes, you can generate batch configurations from plain terms to automate image search across multiple topics. This workflow specifically supports note enrichment for Obsidian notes.

You can generate image search configurations directly from plain terms to automate the fetching and embedding process. This allows you to quickly prepare visuals for articles and presentations.

Using an LLM for image selection automates the ranking of Google Custom Search results, saving time and ensuring consistent quality. It eliminates the inconsistency of manually finding relevant visuals.