Gemini supports a broad set of file types, while exact limits depend on account, plan and resource availability.
1. Add the file from the appropriate source
Use Add files to upload from your device, Drive, a code folder or another supported source available on your account.
2. State what the file is and why it matters
Give the document type, timeframe, owner/version and the specific question you need answered.
3. Ask for structure before interpretation
For complex files, first ask for sections, tables, key entities, missing fields or code layout before asking for conclusions.
4. Separate extraction from analysis
First ask “what does the source say?” Then ask “what does it imply?” so quoted evidence and model interpretation do not blur together.
5. Verify the source location
For important facts, request page/section/table/file-path references and inspect the original material yourself.
Use the uploaded material as the evidence base. First map its structure. Then answer my question with two columns: “directly stated in source” and “your analysis.” Include source locations and list anything you could not verify.
Uploading a large folder without a question; treating file names as proof of version; asking Gemini to infer missing values silently; mixing web facts with file facts without labeling them.
Finish check
A useful file workflow lets another person reproduce the answer by opening the same source and checking the cited location.