Google extended its Gemini API File Search tool with the three things every production RAG team has been begging for: native multimodal indexing, custom metadata filtering at query time, and page-level citations. PDFs, scientific imagery, and plain text now live in one searchable index, powered by Gemini Embedding 2. The post hit the Hacker News front page this week and reads as a direct shot at Anthropic, which still leans on file uploads for enterprise document search inside Claude.
What you get inside the API
File Search is a managed RAG primitive on the Gemini API. Upload a mixed-media collection, attach custom key-value metadata to each file, and query naturally — Google handles chunking, embedding, retrieval, and citation. Every answer comes back with the exact page reference, which is the part legal, medical, and finance teams actually need before they ship anything to users. Billing is per indexed token, with no separate vector DB to run.
Why it lands now
Multimodal plus citations is the painful middle of building verifiable RAG. Most teams stitched together pgvector, an embedding model, and a citation layer themselves. Google just collapsed that stack into one API call, and undercut every enterprise-search startup raising on the same pitch.
You Might Also Like
- Google Releases Gemini Embedding 2 one Vector Space for Text Images Video and Audio
- Google Rolls out File Generation in Gemini Chasing Chatgpt Canvas and Claude Artifacts
- Gemini Canvas in ai Mode Google Just Turned Search Into a Creative Workspace
- Mcp2cli the Tool That Cuts mcp Token Costs by 99 Just hit Hacker News
- Anthropic Just Launched Code Review in Claude Code and 54 of prs now get Real Feedback
Top AI Product 
Leave a comment