Skip to main content

Quick Start Guide

This guide will walk you through the essential steps to get Antigravity Manager up and running, from adding your first account to making your first successful API call.

Prerequisites

Before you begin, ensure you have:

Antigravity Manager Installed

Follow the Installation Guide if you haven’t installed it yet

Google Account

A Google account with access to Gemini API or Google One AI Premium

Step 1: Launch Antigravity Manager

1

Start the Application

  • Desktop: Launch “Antigravity Tools” from your application menu
  • Docker: Navigate to http://localhost:8045 in your browser
2

Set Your API Key

Go to API ProxySettings and set your preferred API key. This key will be used to authenticate all API requests.
# Example API key format
sk-antigravity
Remember this key - you’ll need it for all API calls!
3

Start the Proxy Service

In the API Proxy tab, click the toggle to start the local proxy server. The default address is:
http://127.0.0.1:8045

Step 2: Add Your First Account

Antigravity Manager supports multiple methods for adding accounts. OAuth 2.0 is recommended for the best experience.
1

Navigate to Accounts

Click on Accounts in the sidebar, then click Add AccountOAuth
2

Copy Authorization URL

The dialog will pre-generate an authorization URL. Click the link to copy it to your clipboard.
The authorization URL contains a one-time local callback port. Always use the latest URL shown in the dialog.
3

Complete Authorization

  1. Open the copied URL in your preferred browser
  2. Sign in with your Google account
  3. Grant the requested permissions
  4. Wait for the browser to show ”✅ Authorized successfully!”
4

Finalize in Antigravity

The application should automatically detect the authorization and save your account. If not, click “I already authorized, continue” to finish manually.
If the application isn’t running or the dialog is closed during authorization, the browser may show localhost refused connection. Simply restart the OAuth flow from Step 1.

Alternative: Token Import

If you already have tokens from another tool:
  1. Click Add AccountToken
  2. Paste your token
  3. Click Save

Step 3: Verify Account Status

After adding your account:
1

Check Account List

Your account should appear in the Accounts page with:
  • ✅ Active status
  • Current quota percentage for each model
  • Last synchronization time
2

View Dashboard

Navigate to Dashboard to see:
  • Average remaining quota across all accounts
  • Best account recommendations
  • Active account snapshot
3

Refresh Quota

Click Refresh next to your account to sync the latest quota information from Google.
Antigravity Manager automatically detects and marks accounts with 403 Forbidden status, skipping them in routing.

Step 4: Make Your First API Call

Now that your account is configured, let’s make your first API call. Choose your preferred integration method:

Using Python

The most straightforward way to test the integration:
import openai

client = openai.OpenAI(
    api_key="sk-antigravity",  # Your API key from Step 1
    base_url="http://127.0.0.1:8045/v1"
)

response = client.chat.completions.create(
    model="gemini-3-flash",
    messages=[{"role": "user", "content": "Hello, please introduce yourself"}]
)

print(response.choices[0].message.content)

Using Claude Code CLI

Integrate with the official Claude Code command-line interface:
1

Set Environment Variables

export ANTHROPIC_API_KEY="sk-antigravity"
export ANTHROPIC_BASE_URL="http://127.0.0.1:8045"
2

Run Claude

claude
You should now be able to interact with Claude through Antigravity Manager’s proxy.

Using OpenCode

Antigravity Manager includes built-in OpenCode synchronization:
1

Navigate to OpenCode Sync

Go to API ProxyExternal ProvidersOpenCode Sync
2

Click Sync Button

This automatically generates ~/.config/opencode/opencode.json with:
  • Dedicated provider antigravity-manager (doesn’t overwrite google/anthropic)
  • Optional: Check Sync accounts to export account data
3

Test the Integration

# Test antigravity-manager provider (supports --variant)
opencode run "test" --model antigravity-manager/claude-sonnet-4-5-thinking --variant high

# If opencode-antigravity-auth is installed, google provider still works independently
opencode run "test" --model google/antigravity-claude-sonnet-4-5-thinking --variant max
Windows Users: The config path is C:\Users\<User>\.config\opencode\ following the same ~/.config/opencode convention.

Using cURL

Test the API directly with HTTP requests:
curl -X POST http://127.0.0.1:8045/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-antigravity" \
  -d '{
    "model": "gemini-3-flash",
    "messages": [
      {"role": "user", "content": "What is the capital of France?"}
    ]
  }'

Understanding Model Names

Antigravity Manager provides flexible model naming for different use cases:

Available Models

Model NameDescriptionUse Case
gemini-3-flashFast, efficient modelQuick queries, high-volume
gemini-3-pro-highAdvanced reasoningComplex tasks
gemini-3.1-proLatest Gemini ProProduction workloads
claude-sonnet-4-6Claude 4.6 SonnetBalanced performance
claude-sonnet-4-6-thinkingWith thinking modeComplex reasoning
claude-opus-4-6-thinkingMost capable modelAdvanced tasks

Model Routing

You can configure custom model mappings in Model Router:
  1. Navigate to Model Router in the sidebar
  2. Click Add Mapping
  3. Set source model pattern (regex supported)
  4. Set target model
  5. Save and test
Use regex patterns like gpt-4.* to route all GPT-4 requests to gemini-3-pro-high

Advanced Features

Smart Account Switching

Antigravity Manager automatically selects the best account based on:
  • Current quota availability
  • Account tier (Ultra/Pro/Free)
  • Reset frequency
  • 403 forbidden status
View recommendations in the DashboardBest Account section.

Image Generation Parameters

size
string
Aspect ratio or resolution. Supports:
  • Specific dimensions: 1920x1080, 1024x1024, 1280x720
  • Aspect ratios: 16:9, 9:16, 4:3, 3:4, 1:1, 21:9
System automatically maps to standard ratios.
quality
string
default:"standard"
Image quality level:
  • "hd" → 4K resolution
  • "medium" → 2K resolution
  • "standard" → Default resolution
imageSize
string
Direct Gemini resolution specification (highest priority):
  • "4K" - 4K resolution
  • "2K" - 2K resolution
  • "1K" - Standard resolution
Overrides quality parameter if both are set.

Quota Protection

Enable quota protection to prevent account exhaustion:
  1. Go to SettingsQuota Protection
  2. Enable the toggle
  3. Set minimum quota threshold (e.g., 10%)
  4. Antigravity will skip accounts below the threshold

Background Tasks

Configure automatic account maintenance:
Automatically sync account quotas at regular intervals:
  • SettingsBackground Auto Refresh
  • Set interval (in minutes, max 35791)
  • Recommended: 15-30 minutes
Keep accounts active with periodic requests (currently disabled by default in v4.1.24+):
  • Manual warmup still available in Account Management
  • To re-enable: modify source code and rebuild

Integration Examples

Cherry Studio

  1. Open Cherry Studio settings
  2. Add a new OpenAI provider:
    • Base URL: http://127.0.0.1:8045/v1
    • API Key: sk-antigravity
  3. Configure model settings:
    • Model: gemini-3-pro-image (for images)
    • Size: 1920x1080
    • Quality: hd

Kilo Code

Use Gemini protocol with Kilo Code. OpenAI mode produces non-standard paths (/v1/chat/completions/responses) that return 404.
  1. Open Kilo Code settings
  2. Select Gemini as protocol
  3. Set Base URL: http://127.0.0.1:8045
  4. Set API Key: sk-antigravity
  5. Configure model mappings if needed

Troubleshooting

Check if proxy service is running:
  1. Open Antigravity Manager
  2. Navigate to API Proxy
  3. Ensure the toggle is ON (green)
Verify port isn’t blocked:
curl http://127.0.0.1:8045/health
Verify your API key matches:
  1. Check API ProxySettings for the configured key
  2. Ensure your requests use the same key in headers:
    • OpenAI format: Authorization: Bearer sk-antigravity
    • Anthropic format: x-api-key: sk-antigravity
Possible causes:
  1. Verification required: Check error details for verification link
  2. Plan ineligibility: Verify your Google account has Gemini access
  3. Rate limiting: Wait and try refreshing quota later
Solution:
  • Click on the account to view error details
  • Follow any verification links provided
  • Some 403 errors resolve automatically after waiting
Refresh the quota manually:
  1. Go to Accounts
  2. Click Refresh next to the account
  3. Wait for synchronization to complete
Enable auto-refresh:
  • SettingsBackground Auto Refresh
  • Set interval to 15-30 minutes
Check image model quota:
  1. Verify accounts have gemini-3-pro-image or gemini-3.1-flash-image quota
  2. Dashboard should show non-zero “Gemini Image” average quota
  3. Refresh account quotas if needed
Common issue (fixed in v4.1.27):
  • Ensure using v4.1.27+ which includes gemini-3.1-flash-image support
  • Earlier versions only recognized gemini-3-pro-image

Next Steps

Explore Model Router

Set up custom model mappings and routing rules for advanced workflows

Configure Docker

Deploy Antigravity Manager on your NAS or server for 24/7 availability

Join Community

Get help, share tips, and stay updated on new features

GitHub Repository

Report issues, request features, and contribute to the project

Common Use Cases

Add multiple Google accounts to:
  • Increase total quota pool
  • Automatic failover when one account is rate-limited
  • Smart routing based on quota availability
Antigravity Manager handles rotation automatically.
Use Antigravity Manager as a local proxy for:
  • Testing AI integrations without cloud dependencies
  • Rapid prototyping with multiple model types
  • Cost-effective development with Gemini’s free tier
Deploy with Docker and share:
  • API_KEY with team members for AI requests
  • WEB_PASSWORD with admins only for account management
  • Centralized quota monitoring and account health
You’re all set! You now have Antigravity Manager configured and ready to proxy AI requests. Experiment with different models, set up custom routing, and explore advanced features.

Build docs developers (and LLMs) love