/setup-gemini.en

description: "Lesson command"
duration: "~10 min"
prerequisites: ["Have a Google account", "Browser available"]
level: "beginner"
tags: ["setup", "gemini", "api"]
nonInteractiveMode: incompatible

Gemini API Setup

Step 0: Check Setup Progress

**What the AI auto-runs:** 1. Run `uv run python tools/setup_progress.py show --current setup-gemini` to display progress 2. Auto-detect existing API key:

• Run `uv run python tools/credential_manager.py status`
• If GEMINI_API_KEY is already set, you can just run Step 3 (API test) and mark it as complete
• If a plaintext key exists in `.env`, suggest migrating it to the credential store

What You'll Do in This Session

| Item | Details | |------|---------| | Goal | Get a Gemini API key from Google AI Studio and save it to the Credential Store so you can use AI features like image generation | | Duration | ~10 minutes | | Prerequisites | Have a Google account and a browser available | | Skill Level | No CLI commands needed (everything is auto-run by AI + GUI operations only) |

**Session flow:** 1. Open Google AI Studio in the browser (AI opens the browser automatically) 2. Get your API key (just click buttons on screen) 3. Save the API key securely (AI creates the file automatically) 4. Enter the API key (open the file and paste) 5. Functionality test (AI runs automatically)

> **Hint**: If the AI's response stops midway, type "please continue" or "it stopped" to resume.

---

Readiness Check

**AskQuestion configuration:**

{
  "title": "Pre-session confirmation",
  "questions": [{
    "id": "readiness",
    "prompt": "Are you ready?",
    "options": [
      {"id": "ready", "label": "Ready! Let's start"},
      {"id": "chrome", "label": "Automate browser operations with /chrome"},
      {"id": "check_prereq", "label": "I want to check the prerequisites"},
      {"id": "different_lesson", "label": "I want to go to a different lesson"}
    ]
  }]
}

(ready -> Proceed to Step 1) (chrome -> After opening the browser in Step 1, follow the "Automating with Chrome integration" section for automatic execution) (check_prereq -> Guide: "You're ready if you can log in to a browser with a Google account.") (different_lesson -> Display module list)

---

Step 1: Open Google AI Studio in the Browser

**What the AI does:** 1. Auto-detect the OS (Mac / Windows / Linux) 2. Run the following command to open the browser automatically:

# Mac:
open https://aistudio.google.com/apikey
# Windows:
start https://aistudio.google.com/apikey
# Linux:
xdg-open https://aistudio.google.com/apikey

**After the browser opens, display the following AskQuestion:**

{
  "title": "Step 1: Get your API key in the browser",
  "questions": [{
    "id": "browser_status",
    "prompt": "Did the browser open? Follow these steps to get your API key:\n\n1. Log in with your Google account\n2. Click the 'Get API key' button\n3. Click 'Create API key'\n4. Click the 'Copy' button next to the displayed API key\n\nDid you copy the API key?",
    "options": [
      {"id": "copied", "label": "I copied the API key!"},
      {"id": "browser_not_open", "label": "The browser didn't open"},
      {"id": "no_button", "label": "I can't find the 'Get API key' button"},
      {"id": "login_issue", "label": "I can't log in to my Google account"}
    ]
  }]
}

(copied -> Proceed to Step 2) (browser_not_open -> Guide: "Open this URL directly in your browser: https://aistudio.google.com/apikey") (no_button -> Guide: "Wait for the page to fully load. If it still doesn't appear, click the 'Get API key' tab at the top of the page.") (login_issue -> Guide: "Google AI Studio requires a gmail.com or Google Workspace account. Try logging in with your company account.")

---

Automating with Chrome Integration (`/chrome` mode)

**Prerequisite:** The "Claude in Chrome" extension (v1.0.36+) must be installed in Chrome, and you must have launched with `claude --chrome` or run `/chrome` in the session.

**What the AI auto-runs with Chrome integration:** 1. Open https://aistudio.google.com/apikey in the browser 2. Use Chrome integration to perform the following operations in order:

• If Google account login is needed, wait for user's action
• Click the "Get API key" or "Get API key" button
• Click the "Create API key" or "Create API key" button
• If a project selection screen appears, select the default project and click "Create API key in existing project"

3. Once the API key appears on screen, tell the user "Click the copy button to copy the API key" 4. Proceed to Step 2

**Note:** Do not read the API key value from the browser screen. The user copies it manually.

If Chrome integration is not available, follow the Step 1 instructions manually.

---

Step 2: Save the API Key Securely

**Important security note:** Do not paste the API key in this chat. We'll save it securely in a separate terminal window.

**What the AI auto-runs:** 1. Check if the `keyring` package is installed

• If not installed: auto-run `uv add keyring`

2. Run `uv run python tools/credential_manager.py status` to check the current state

**Message to display to the user:**

Once you've copied the API key, follow these steps to save it securely:

┌─────────────────────────────────────────────────────────────┐
│ Run the following command in a separate terminal window:     │
│                                                             │
│ Cursor: Ctrl+` (backtick) to open a new terminal           │
│ Claude Code: Open a separate terminal window                │
│                                                             │
│ uv run python tools/credential_manager.py store GEMINI_API_KEY     │
│                                                             │
│ → "Enter value for GEMINI_API_KEY:" will be displayed       │
│ → Paste the copied API key and press Enter                  │
│   (The characters you type won't be shown on screen.        │
│    This is nor