Gemini CLI - first steps after installation

This article covers what happens on first run how authentication works and how to start using Gemini CLI right away.

Initial setup wizard

When you run Gemini CLI for the first time an interactive setup wizard starts automatically.

Step 1 – Launch Gemini CLI

Start the CLI by running:

gemini

The wizard guides you through the initial configuration using keyboard navigation.

Step 2 – Select a visual theme

You’ll be prompted to choose a visual theme. This controls colors and layout in the terminal interface.

Use the arrow keys to browse options and press Enter to confirm your choice. This setting only affects display and can be changed later.

Authentication methods

After choosing a theme Gemini CLI asks how you want to authenticate. Authentication is required before sending any requests.

Step 3 – Choose a login method

The most common option is signing in with a personal Google account. This provides access to the free usage tier with defined rate limits.

Other supported options include:

Google AI Studio API key set via the GEMINI_API_KEY environment variable
Google Cloud Vertex AI using GEMINI_API_KEY and GOOGLE_GENAI_USE_VERTEXAI=true

These options are useful for higher limits or enterprise environments.

Step 4 – Complete browser authentication

After selecting the method a browser window opens or a login URL is shown in the terminal.

Sign in securely and return to the terminal once finished. A confirmation message indicates that authentication is complete and the CLI is ready.

Using Gemini CLI after setup

Once configuration and authentication are done you can start using Gemini CLI immediately.

You can run Gemini in any directory. The CLI automatically works in the context of the current folder and its files.

Common usage patterns

Start working in a project by navigating to its directory and running gemini
Ask questions or request code using natural language
Reference local files in prompts using @filename
Explore available commands by typing / inside the CLI

Each prompt is evaluated using the files and structure of the current directory which makes Gemini useful for project-specific tasks.

Helpful tips and troubleshooting

Most issues after setup are easy to resolve.

Authentication problems

If login fails check that your browser is signed into the intended Google account before launching Gemini CLI again.

Node.js version warnings

If you see version errors make sure the system meets the minimum Node.js requirement used by the template.

Network or model fallback

On unstable connections Gemini CLI can switch models automatically to reduce latency and continue responding.

Advanced usage on Ubuntu

Gemini CLI works well inside editor terminals such as Visual Studio Code. Open an integrated terminal and run gemini to use it as part of your daily development flow.

You can also reference files mid-conversation using @ to review edit or analyze them in context.

For advanced integrations MCP servers allow Gemini CLI to interact with other tools or internal systems.

Authentication limits and plans

Usage limits depend on the authentication method used.

Personal Google account access includes daily and per-minute request limits
Google AI Studio API keys provide a small free tier with paid upgrades available
Vertex AI access allows higher limits when linked to a billing account

Review limits carefully if you plan to use Gemini CLI for production workloads or automation.

Community and documentation

Gemini CLI is open source and actively developed. Bugs feature requests and improvements are handled openly by the community.

For detailed flags advanced authentication options and updates refer to the official Gemini CLI documentation and repository.