Complete guide to Gemini
Getting started
Go to gemini.google.com and sign in with your Google account; the free plan already gives you access to most basic features. Type your question or upload a document, image, or file for it to analyze. For complex questions that require in-depth research, select "Deep Research" in the prompt bar: Gemini turns your request into a multi-point research plan, browses the web (and your Gmail/Drive/Chat if you authorize it), and delivers a complete report, even available as an audio summary.
Use cases
Automated deep research
Deep Research breaks a complex question down into sub-tasks, searches and browses the web autonomously, and synthesizes a complete report with sources.
Specialized assistants with Gems
Create or use predefined Gems (hiring consultant, copywriter, sentiment analyzer) that apply specific instructions only when you open them, without affecting your other conversations.
Google ecosystem integration
Analyze and act on content from Gmail, Drive, and Chat directly from the chat, using the context of your Google account.
High-volume multimodal analysis
Thanks to its context window of over 1M tokens, it can analyze very long documents or multiple files at once without losing coherence.
Advanced setup
- Enable Deep ResearchSelect it directly in the prompt bar before typing your question. You can review and adjust the research plan before Gemini runs it, and you can also upload your own files as sources.
- Create a GemOpen the menu and go to "Gems" to access the manager: you can use Google's predefined Gems, your previous Gems, or create a new one by uploading files, connecting Drive, or adding NotebookLM notebooks.
- Available plansThe free plan gives limited access to Deep Research (available on the 2.5 Flash model at no cost). Google AI Pro ($19.99/month) and Ultra extend access to Gemini 3.1 Pro and additional professional tools such as NotebookLM and Flow.
- Difference between Gems and custom instructionsCustom instructions apply to all conversations; Gems only activate when you open that specific Gem, so you can run several specialized assistants in parallel.
Tips
- Use Deep Research selectively, for questions that genuinely require researching multiple sources, not for simple queries: it takes longer than a normal response.
- Review and adjust the research plan before it runs: you can redirect the focus if you see it's going to concentrate on something you're not interested in.
- Create one Gem for each type of recurring task instead of repeating instructions every time: it keeps each assistant specialized and consistent.
- Take advantage of the Gmail and Drive integration for questions about your own content, instead of searching manually through emails or documents.
- For very long documents, upload the entire file instead of fragments: Gemini's large context window is designed for that.
FAQ
Is Gemini free to use?
Yes, it has a functional free plan with limited access to Deep Research. Google AI Pro ($19.99/month) and Ultra extend access to the most powerful models and additional tools.
What exactly is a Gem?
It's a custom assistant inside Gemini with its own instructions and files, which only activates when you open it, unlike custom instructions, which always apply.
Does Deep Research access my emails and files?
Yes, if you authorize it, it can browse your Gmail, Drive, and Chat in addition to the public web to build the report.
Is Gemini integrated into Google Workspace?
Yes, it's available for Google Workspace business and education plans, with access to additional professional tools.
Based on gemini.google/release-notes and official Google DeepMind documentation