antigravity-studio

Google Antigravity (Cloud Code Assist) client in pure Python — text chat + image generation + agentic tasks.

Enhanced with strategies from Antigravity-Manager Rust project.

Key features: 3-tier endpoint fallback (Sandbox → Daily → Prod), free-tier project ID fallback, safety settings OFF, auto-failover across accounts, mandatory cooldown after image generation.

Install

1. Global CLI Access (NPM)

Install the studio globally to access the command from anywhere:

npm install -g @comgunner/antigravity-studio

2. Setup Python Environment

Since the core logic is Python-based, you need to set up the environment in the package location:

# Navigate to the installed package (or your cloned repo)
cd $(antigravity-studio path | xargs dirname)

# Create and activate virtual environment
python3.12 -m venv .venv
source .venv/bin/activate

# Update base tools and install dependencies
python -m pip install --upgrade pip setuptools wheel
pip install -r requirements.txt

Picoclaw-Agents Integration (As a Skill)

If you use picoclaw-agents, you can install this studio as a Skill automatically:

# Install as a Skill for Picoclaw
antigravity-studio install --target picoclaw

# Other targets: --target claude, --target agents

Quick Start

1. Login

# Browser OAuth (recommended)
python3 antigravity_cli.py login

# Login to a specific account
python3 antigravity_cli.py login --account work

# Headless / SSH / Termux
python3 antigravity_cli.py login --device

1b. Multi-Account Setup

# Add a second account
python3 antigravity_cli.py accounts add work

# Login to it (opens browser OAuth)
python3 antigravity_cli.py login --account work

# List all accounts (shows auth status, emails, project IDs)
python3 antigravity_cli.py accounts

# Switch active account
python3 antigravity_cli.py accounts switch work

2. List Models

python3 antigravity_cli.py models

Output Example:

Available Antigravity Models:
────────────────────────────────────────────────────────────
  ✓ claude-opus-4-6-thinking (Claude Opus 4.6 (Thinking))
  ✓ claude-sonnet-4-6 (Claude Sonnet 4.6 (Thinking))
  ✓ gemini-2.5-flash (Gemini 3.1 Flash Lite)
  ✓ gemini-2.5-flash-lite (Gemini 3.1 Flash Lite)
  ✓ gemini-2.5-flash-thinking (Gemini 3.1 Flash Lite)
  ✓ gemini-2.5-pro (Gemini 2.5 Pro)
  ✓ gemini-3-flash (Gemini 3 Flash)
  ✓ gemini-3-flash-agent (Gemini 3 Flash)
  ✓ gemini-3-pro-high (Gemini 3 Pro (High))
  ✓ gemini-3-pro-low (Gemini 3 Pro (Low))
  ✓ gemini-3.1-flash-image (Gemini 3.1 Flash Image)
  ✓ gemini-3.1-flash-lite (Gemini 3.1 Flash Lite)
  ✓ gemini-3.1-pro-high (Gemini 3.1 Pro (High))
  ✓ gemini-3.1-pro-low (Gemini 3.1 Pro (Low))
  ✓ gpt-oss-120b-medium (GPT-OSS 120B (Medium))

3. Chat

python3 antigravity_cli.py chat "What is the capital of France?"
python3 antigravity_cli.py chat "Write a poem" --model gemini-3-flash

Using gemini-3-flash-agent

The gemini-3-flash-agent model supports agentic capabilities — it can use tools, execute code, and perform multi-step reasoning:

# Basic agent chat
python3 antigravity_cli.py chat "Analyze this Python code and suggest improvements" --model gemini-3-flash-agent

# Agent with a complex task
python3 antigravity_cli.py chat "Write a function to sort a list, then test it with 5 edge cases" --model gemini-3-flash-agent

# Agent with extended context
python3 antigravity_cli.py chat "Debug this error: IndexError: list index out of range" \
  --model gemini-3-flash-agent --max-tokens 8192

4. Technical Summary (Multi-Asset Analysis)

Generate technical summaries for Crypto (Binance), Metals, Forex, and Indices (Yahoo Finance). This feature calculates EMAs (3, 9, 21, 50, 200) and uses Gemini for narrative technical analysis.

Supported Assets (Yahoo Finance):

• Metals: xau (Gold), xag (Silver)
• Forex: eurusd, gbpusd, jpymxn, mxn
• Indices: gspc (S&P 500), ixic (Nasdaq), dxy (Dollar Index)
• Commodities: cl (Crude Oil)
• Crypto (Binance): btc, eth, sol, ada, etc.

# 1. Standard Crypto (Binance)
python3 antigravity_cli.py --resume btc

# 2. Gold (Yahoo: Gold Futures GC=F)
python3 antigravity_cli.py --resume xau

# 3. Nasdaq (^IXIC)
python3 antigravity_cli.py --resume ixic --tf 1h

Example Output (BTC):

--- Generating BTC 4h Summary with Gemini ---
CURRENT STATUS: BTC (4h)
Price: $74,520.3 (0.56%)
EMA 3:   74210.15 | EMA 9:   73990.86
EMA 21:  73158.25 | EMA 50:  71894.85 | EMA 200: 69932.71

--- AI TECHNICAL VIEW (GEMINI) ---
**Sentiment: Strongly Bullish**, as price holds above all major EMAs (3, 9, 21, 50, 200) in a clear technical uptrend.

Example Output (Gold):

--- Generating XAU 1h Summary with Gemini ---
CURRENT STATUS: XAU (1h) | SOURCE: Yahoo Finance
Price: $4,872.8 (0.27%)
EMA 3:   4866.66 | EMA 9:   4856.08
EMA 21:  4833.55 | EMA 50:  4804.68 | EMA 200: 4748.71

Example Output (Crude Oil):

--- Generating CL 1h Summary with Gemini ---
CURRENT STATUS: CL (1h) | SOURCE: Yahoo Finance
Price: $90.15 (-0.23%)
EMA 3:   90.44 | EMA 9:   91.57
EMA 21:  93.59 | EMA 50:  96.28 | EMA 200: 99.74

5. Generate Images

Simple Prompt

python3 antigravity_cli.py img "A cute cat wearing sunglasses" -o cat.png

Aspect Ratio & High Detail

python3 antigravity_cli.py img "Neo from the Matrix... falling code" --aspect-ratio 4:5 -o neo.png

Cinematic Portait

python3 antigravity_cli.py img "Portrait of a warrior" --aspect-ratio 9:16 -o warrior.png

Advanced: JSON Prompt with Reference Image

python3 antigravity_cli.py img '{                                                         
    "id": 1,
    "theme": "Python Coding",
    "prompt": "A promotional graphic with a dark high-tech background. In the center, a 3D stylized Python logo with a glowing blue and yellow circuit board pattern, digital steam rising. Top center: include the logo from the reference image. Bottom text in bold white sans-serif: Python without Errors. Minimalist, clean, high quality 3D render."
}' -r ./logo/logo_python.png -o python_no_errors.png

Project Structure

| File | Purpose | |------|---------| | antigravity_cli.py | Main entry point for all commands | | coin_summary.py | Multi-asset technical analysis logic | | bin/cli.js | NPM Global Installer logic | | SKILL.md | Agent capability manifest |

License

This project is licensed under the GNU General Public License v3.0. See the LICENSE.md file for details.