Skip to main content

​
Getting started

Do I need to know how to code?

No. You can create an AI trader, connect an exchange wallet, and deploy a prompt strategy entirely through the web app without writing a line of code. Program strategies require a small amount of Python, but the built-in starter templates are enough for most users.
Hyperliquid and Binance Futures are supported for live trading. Both offer testnet environments, and we strongly recommend starting there.
Yes. You bring your own key for the language model your trader uses — OpenAI, Anthropic, Google, DeepSeek, or any OpenAI-compatible endpoint. This keeps LLM costs predictable and under your control.
Hyperoru has no minimum. Your exchange does. On mainnet, both Hyperliquid and Binance Futures accept small positions, though fees and funding start to matter at very small sizes. On testnet you can experiment for free.
Yes. Each AI trader is fully isolated: its own keys, its own strategy, its own on/off switch, its own analytics. Many users run a handful of traders in parallel with different strategies.

​
Strategies

Start with a prompt strategy if you think in terms of trading ideas, market narratives, or human-readable rules. Pick a program strategy if your rules are precise (e.g. “buy when 20-EMA crosses 50-EMA on the 1h chart, size = 1% of equity per trade”) and you want deterministic, repeatable behavior.You can run both side by side and compare.
Only when a signal fires. If you set up a breakout signal that fires twice a day, your strategy runs twice a day. If you set up a high-frequency signal, it runs more often. Strategies never run continuously by themselves.
Yes. Both prompt and program strategies can be run against historical data from the Backtests panel. See Backtesting.
Prompt strategy runs have a timeout. If the model times out or returns an invalid decision, the run is recorded as an error and nothing is traded. Program strategies do not depend on an external LLM.
Yes. Edits take effect on the next decision run. If you are making a large change, it is safer to stop the trader, verify the new strategy in a backtest, then restart.

​
Money and risk

No. Your funds live at the exchange, under your own account. Hyperoru only holds keys that let it place trading orders. See Trust and safety.
Your positions continue to live at the exchange, whether or not Hyperoru is reachable. Stop losses and take profits that Hyperoru placed on the exchange side stay active. You can always log in to the exchange directly to manage positions.
Yes. In each trader’s risk settings you can cap maximum notional, maximum leverage, maximum positions, and daily loss limits. When a cap is hit, the trader stops opening new positions.
In the app, toggle Stop Trading on the trader. Through the API, call POST /api/account/{account_id}/disable-trading. Either halts new orders in seconds.
The only fees you pay are (1) your exchange’s trading fees, (2) your LLM provider’s usage, and (3) the Hyperoru subscription you signed up for. Hyperoru does not take a cut of your trades.

​
Data and privacy

No. Your prompts, programs, traders, and trade history are private to your account. We use aggregated, anonymized metrics to improve the platform, never individual strategies.
Yes — when a prompt strategy runs, the assembled market context and your instructions are sent to the LLM you chose. Check the privacy policy of your LLM provider for their retention policy. Program strategies never touch an external LLM.
Yes. Email support@hyperoru.com to request a full export of your account, traders, decisions, and trades. Programmatic export is also possible via the API.
Email support@hyperoru.com. Make sure you have stopped all traders and withdrawn funds from connected exchanges first.

​
API

Log in with POST /api/users/login to get a session token, then send it in the Authorization: Bearer header on every request. See Authentication.
120 reads per minute per token, 60 writes per minute per token, and a few per-endpoint concurrency limits. See Errors and rate limits.
Not officially, but the API follows standard OpenAPI 3.1. Generate a typed client in your language of choice from the spec. Many open-source generators (openapi-generator, nswag, oapi-codegen) produce a usable client in minutes.
Yes. Open a WebSocket to wss://api.production.hyperoru.com/ws?session_token=... and subscribe to the channels you care about. See WebSocket.

​
Support

Email support@hyperoru.com with the request id from the error (if any), the approximate UTC timestamp, and a short description. We aim to respond within one business day.
Check status.hyperoru.com first. If the incident is not listed, email support.
Email support@hyperoru.com with “Feature request” in the subject. We read every one.