To ask questions of your content once, skip the index and use Gemini 3.8 Flash.

Why Gemini 3.8 Flash

The reasons it wins for this job.

Building a search index is a recurring-job solution. For one set of questions, a 1M-token context window holds hundreds of pages directly.
Gemini 3.8 Flash takes text, images, video, audio and PDF as input, so a folder of mixed material goes in as-is.
It ranks level with Claude Opus 5 on LMArena's text leaderboard, from the low cost tier.
In the Gemini app it works alongside Google Drive, where much of this content already lives.

How to use it

4 steps to a first result.

1Collect the materialExport the docs, slides and recordings that could contain the answer. Leave out the rest.
2Upload in one conversationEverything in one place, so it can connect facts across files.
3Ask for the source per answerFile name and page or timestamp for every claim.
4Check the ones you act onOpen the cited file. If it cannot cite, treat the answer as a guess.

Prompt for Gemini

Start from this.

Edit the parts in capitals, then run it.

The attached files are WHAT THEY ARE. Answer my questions using only these files.

For every answer:
- Cite the file name and the page, slide or timestamp
- If files disagree, show both and say which is newer
- If the answer is not in the files, say "not in the files"

First question: QUESTION

Nothing is sent anywhere. It copies to your clipboard.

Alternatives that also work

If Gemini 3.8 Flash is not an option.

Claude Opus 5Official page →

Anthropic · closed · cost: high

Pick it when the content is text-heavy and the answer needs careful reasoning across documents.

GPT-6 AstraOfficial page →

OpenAI · closed · cost: high

Pick it when your files are in ChatGPT already and the question mixes your content with web research.

Qwen3.8 27BHugging Face →

Alibaba Qwen · open weights · cost: free weights; you pay for the hardware · 7.5M HF downloads/mo

Pick it when the material is confidential. It reads text and images, runs on one GPU, and extends to a 1M context.

Watch out for

Long context is not free recall. Facts buried in the middle of very long inputs are missed more often. Ask narrow questions.
If you ask the same kind of question every week, move to the recurring setup and stop re-uploading.

Doing this again and again? For search over your content, embed it with Gemini Embedding 2. A help-centre assistant or internal search runs thousands of queries against content that changes. That needs embeddings. Gemini Embedding 2 embeds text, images, video, audio and PDF into one space, with adjustable dimensions and a low price per token.

Sources

Checked 17 September 2026. Models change monthly; we re-check this page when they do.

Picking the model is the easy part.

Wiring it into a workflow that runs every week, with evals, fallbacks and a cost you can predict, is the work. Fifteen minutes, no deck.

Book fifteen minutes →