Your own logic
Use pandas, NumPy or your own models. Any PyPI package with a binary wheel installs from requirements.txt, and npm packages from package.json.
Python · Node.js · Hosted trading bots
Host your Python
trading bot on Hyperliquid.
Antei hosts Python and Node.js trading bots on Hyperliquid. Upload a file, a folder or a zip with your requirements.txt or package.json, and Antei builds an immutable version in an isolated sandbox, backtests it on historical market data and runs it in paper trading or live. Orders go through Antei's order system, so risk limits apply outside your code.
01 · How it works
How Python works on Antei
Create your account
Sign up at antei.ai and open Python under Build in your dashboard.
Upload your code
Add a main.py or main.js, or a whole folder or zip, with requirements.txt or package.json. Antei's SDK is already installed, so you never add it yourself.
Build a version
Antei scans the upload for malware and builds an immutable version. Build logs, the markets and timeframe it detects, and any missing packages show next to it.
Backtest it
Replay a market over up to two years of history. Backtests run without internet access, so a test can never reach a live service.
Deploy with limits
Prepare a deployment in paper trading or live with allowed markets, capital, stop loss, take profit, maximum drawdown, daily loss, position size and leverage.
02 · Capabilities
What you can do with Python
Antei SDK
The Python and JavaScript SDKs place market, limit and post-only orders through Antei's order system and report every order's status and fills.
Immutable versions
Every upload becomes a version you can inspect, backtest and deploy, so a running bot never changes underneath you.
Build and run logs
Follow the build while it runs, then the strategy's own logs, orders and positions once it trades.
Risk limits you can change
Adjust capital, loss, position and leverage limits on a running deployment without uploading again.
Deploy from your tools
Upload from the dashboard, from Claude Code or Codex over MCP, or from your own scripts and CI with an API key.
03 · Safety
Isolated by design
- Each bot runs in its own gVisor sandbox with a read-only file system and no extra Linux capabilities.
- Network access is limited to the domains your strategy declares; orders go through Antei, never straight to Hyperliquid.
- Your exchange keys never enter the sandbox: Antei signs orders outside your code.
- Uploads are scanned for malware before they build.
04 · FAQ
Python questions, answered
Can I host a Python trading bot for Hyperliquid on Antei?
Yes. Upload your Python code with a requirements.txt, and Antei builds it, backtests it and runs it around the clock on Hyperliquid in paper trading or live. Orders go through Antei's SDK, so risk limits and the kill switch apply to every one.
Which Python version and packages can I use?
Python 3.12, with Antei's SDK preinstalled and main.py as the entry point. Packages from PyPI install from requirements.txt as long as they ship binary wheels; links to URLs, Git repositories and editable installs are refused.
Does Antei support JavaScript and Node.js bots?
Yes. Upload a main.js with a package.json and its lock file, and Antei builds it on Node.js 22 with its JavaScript SDK. Packages install from the npm registry.
Can my bot call an AI model or another API?
Yes, for the domains your strategy declares; all other network access is blocked. Strategies that call AI models or outside APIs are not backtestable by default, because backtests run without internet access.
How do I stop a running bot?
Pause or kill it from Deployments in your dashboard, or ask your AI assistant over MCP. Kill cancels its resting orders, closes its positions and stops the strategy.
Start with Python