Skip to main content

Installing GenAI Command Line Interfaces on a Win/Mac

CLI Tooling Setup

Currently, Cloud Services supports three models for CLI access as outlined below. Our preferred configuration is leveraging Gemini CLI (with Claude CLI available under certain conditions).

  • Option 1 can be a quick and easy setup for a single user leveraging their Google AI Pro license.
  • Option 2 provides a proxy interface that allows you to manage API tokens (with administrative privileges to create additional tokens and set limits).
  • Option 3 is a more complex offer that links the CLI to a dedicated GCP Project and is for advanced users.

Prerequisites

Local Administrative Access

A Unix-Type Terminal. (Shell)

A working installation of your preferred command line tool:

Access to the AI-CLI Gitlab Repository:

Valid Access to an OSU-Approved GenAI Vendor:

Mac

Option 1 – Gemini CLI using a Google AI Pro license.

user.1@macbook-pro ~ %  ./mac_setup_env_google_ai_pro.sh
Backing up existing  ~/.gemini/.env  to ~/.gemini/.env.bak.20260508_171627

--- Current .env contents (review and merge any custom settings) ---
GOOGLE_GEMINI_BASE_URL=https://litellm.cloud.osu.edu
GEMINI_API_KEY=YOUR_API_KEY_HERE
GOOGLE_CLOUD_PROJECT=osu-prd-geminipro
--------------------------------------------------------------------

Updating existing 'GOOGLE_CLOUD_PROJECT' in ~/.gemini/.env.
Successfully set GOOGLE_CLOUD_PROJECT=osu-prd-geminipro in ~/.gemini/.env.
  • After logging into Gemini from a terminal or vscode, enter the command ‘/auth‘ and choose the ‘Sign in with Google’ option. Following the prompts will open a local browser. Login with your name.#@osu.edu credentials.
  • Go back to Gemini and this time enter the command ‘/about’. You should see:
About Gemini CLI
                                                                                                                                    
 CLI Version			0.37.1 
 Git Commit			3b2d4f100   
 Model				Auto (Gemini 2.5) 
 Sandbox			no sandbox 
 OS				darwin 
 Auth Method			Signed in with Google (name.#@osu.edu) 
 Tier				Gemini Code Assist 
 GCP Project			osu-your-geminipro-project
  • If the Tier is anything other than Gemini Code Assist, stop and contact us as you’re not using a protected endpoint.

Option 2 - CLI Using OTDI Gen AI Virtual Keys (Claude/Gemini)

Claude

user.1@macbook-pro ~ %  .\ mac_setup_claude_settings.sh

Backed up existing settings to ~/.claude/settings.json.bak
Wrote settings to ~/.claude/settings.json
Remember to replace 'YOUR_GEN_AI_API_TOKEN' with your actual token.
  • Use your favorite text editor to open ~/.claude/settings.json and replace 'YOUR_GEN_AI_API_TOKEN' value with the Claude token you created in LiteLLM.

Gemini

user.1@macbook-pro ~ % ./mac_setup_env_gen_ai.sh 
Backing up existing ~/.gemini/.env  to  ~/.gemini/.env.bak.20260508_164058

--- Current .env contents (review and merge any custom settings) ---
GOOGLE_GEMINI_BASE_URL=https://litellm.cloud.osu.edu
GEMINI_API_KEY=[YOUR_API_KEY]
--------------------------------------------------------------------
Your previous .env has been backed up. Review it above and merge any custom settings.

Creating ~/.gemini/.env with placeholder values.
Successfully created ~/.gemini/.env.

IMPORTANT: Open ~/.gemini/.env' and replace the placeholder value for GEMINI_API_KEY with your actual API key.
  • Use your favorite text editor to open ~/.gemini/env and replace the ‘GEMINI_API_KEY‘ value with the Gemini token you created in LiteLLM.

Option 3 - Gemini CLI with a Google GCP API Key (Vertex AI or AI Studio)

  • This setup is defined at Gemini CLI authentication setup.
  • If you would like to use API Tokens from your own GCP account, you can create API tokens in your GCP Project, under the Vertex AI or AI Studio services. This setup will leverage the built in AI services in Google.
  • Request a Project at Ohio State Google Cloud Platform.

Windows 11

Option 1 – Gemini CLI using a Google AI Pro license.

  • From the AI-CLI Gitlab Repository, download & run the following PowerShell script as an Administrator.
    • .\ Win11_setup_env_google_ai_pro.ps1
PS C:\Users\User.1.OSUAD> .\Desktop\Win11_setup_env_google_ai_pro.ps1
Backing up existing C:\Users\User.1.OSUAD\.gemini\.env to C:\Users\User.1.OSUAD\.gemini\.env.bak.20260508_171627

--- Current .env contents (review and merge any custom settings) ---
GOOGLE_GEMINI_BASE_URL=https://litellm.cloud.osu.edu
GEMINI_API_KEY=YOUR_API_KEY_HERE
GOOGLE_CLOUD_PROJECT=osu-prd-geminipro
--------------------------------------------------------------------

Updating existing 'GOOGLE_CLOUD_PROJECT' in C:\Users\User.1.OSUAD\.gemini\.env.
Successfully set GOOGLE_CLOUD_PROJECT=osu-prd-geminipro in C:\Users\User.1.OSUAD\.gemini\.env.
  • After logging into Gemini from a terminal or vscode, enter the command ‘/auth‘ and choose the ‘Sign in with Google’ option.  Follow the prompts which will open a local browser. Login with your Name.#@osu.edu credentials.
  • Go back to Gemini and this time enter the command ‘/about’. You should see:
About Gemini CLI 

CLI Version		0.37.1 
Git Commit		3b2d4f100   
Model			Auto (Gemini 2.5) 
Sandbox			no sandbox 
OS			win32 
Auth Method		Signed in with Google (name.#@osu.edu) 
Tier			Gemini Code Assist 
GCP Project		osu-your-geminipro-project
  • If the Tier says anything other than Gemini Code Assist, stop and contact us as you’re not using a protected endpoint.

Option 2 – CLI Using OTDI Gen AI Virtual Keys (Claude/Gemini)

Claude

  • From the AI-CLI Gitlab Repository, download & run the following script in a PowerShell terminal with Administrator access.
    • .\win11_setup_claude_settings.ps1
PS C:\Users\User.1> .\win11_setup_claude_settings.ps1

Backed up existing settings to C:\Users\User.1.OSUAD\.claude\settings.json.bak
Wrote settings to C:\Users\User.1.OSUAD\.claude\settings.json
Remember to replace 'YOUR_GEN_AI_API_TOKEN' with your actual token.
  • Use your favorite text editor to open C:\Users\User.1.OSUAD\.claude\settings.json and replace 'YOUR_GEN_AI_API_TOKEN‘ with your Claude virtual key from LiteLLM.
  • Once started, you can verify your Claude-CLI setup with the ‘/config’ command, choosing the ‘Status’ menu item.
Settings	   Status	Config	    Usage	Stats

  Version:		2.1.139
  Session name:		/rename to add a name
  Session ID:		11c1fd7d-898a-4838-9001-7fb964543e6e
  cwd:			C:\Users\user.1.OSUAD\Documents
  Auth token:		ANTHROPIC_AUTH_TOKEN
  Anthropic base URL:	https://litellm.cloud.osu.edu
  Model:		Default (claude-opus-4-6[1m])
  Setting sources:	User settings

Gemini

  • From the AI-CLI Gitlab Repository, download & run the following script in a PowerShell terminal with Administrator access.
    • .\win11_setup_env_gen_ai.ps1
PS C:\Users\User.1.OSUAD> .\win11_setup_env_gen_ai.ps1
Backing up existing C:\Users\User.1.OSUAD\.gemini\.env to C:\Users\User.1.OSUAD\.gemini\.env.bak.20260508_164058

--- Current .env contents (review and merge any custom settings) ---
GOOGLE_GEMINI_BASE_URL=https://litellm.cloud.osu.edu
GEMINI_API_KEY=[YOUR_API_KEY]
--------------------------------------------------------------------
Your previous .env has been backed up. Review it above and merge any custom settings.

Creating C:\Users\User.1.OSUAD\.gemini\.env with placeholder values.
Successfully created C:\Users\User.1.OSUAD\.gemini\.env.

IMPORTANT: Open 'C:\Users\User.1.OSUAD\.gemini\.env' and replace the placeholder value for GEMINI_API_KEY with your actual API key.
  • Use your favorite text editor to open C:\Users\User.1.OSUAD\.gemini\.env and replace 'YOUR_API_Key‘ with your Gemini virtual key from LiteLLM.
  • Launch Gemini and enter the command ‘/about’ to confirm your setup:
About Gemini CLI 

 CLI Version	0.25.2
 Git Commit	18e854c33
 Model		auto-gemini-2.5
 Sandbox	no sandbox
 OS		win32
 Auth Method	gemini-api-key
 GCP Project	osu-prd-geminipro
  • There should not be a ‘Tier setting in this config. If there is, please contact us

Option 3 – Gemini CLI with a Google GCP API Key (Vertex AI or AI Studio)

Help

If you find you need help with these installations, please reach out. You can submit a support request through the university’s Help Desk, or book some face time using Ohio State’s GCP Support Bookings Pages.