Crypto is fun, but gambling on price direction isn't really my thing. I wanted to build something more reliable: Funding Rate arbitrage.
Right now, Hyperliquid is a great playground for this: zero fees on the Spot market and deep liquidity on Perpetuals. The plan was simple: write a Python bot to automate a carry trade strategy.
But diving into their API and SDK was... a bit of a ride.
The Strategy: Delta-Neutral Carry Trade
The concept is straightforward. When the market is bullish, Long traders pay a "Funding Rate" to Short traders on perpetual contracts.
The algorithm detects this yield: it shorts the Perp to collect the funding, and simultaneously buys the exact same amount of the Spot asset to hedge. Result: Delta = 0.
You don't care if the market pumps 20% or dumps 20%, your exposure is flat and you just collect the hourly fee passively.
The API Struggle
Wiring this up properly actually took some work. I ran into a few annoying roadblocks:
-
The "U" token trap: In the Hyperliquid ecosystem, Bitcoin is called
BTCon the Perp market, butUBTCon Spot. The official SDK doesn't natively map matching pairs. I had to build a dynamic resolver to auto-map the 280+ assets without hardcoding tickers. -
Async everywhere: On a decentralized orderbook, speed is everything. If you send synchronous API requests, slippage will eat your edge before the hedge even fills. I had to wrap the core execution in
asyncioto fire both legs of the trade at the exact same time. - The execution loop: Knowing when to enter is easy, but exiting when the funding drops, scaling capital dynamically, and handling the overall state requires a robust backend loop.
The Result
Fast forward a few dozen hours, and I finally packaged a clean, working framework.
The bot runs in the background, scans the overlap between Spot and Perp markets, handles the risk math, and runs a local aiohttp web dashboard so I can monitor my PnL directly in the browser.
If you're a developer or a Python tinkerer looking to build on Hyperliquid, I decided to share my full source code. It’ll save you from spending a week fighting the API quirks and writing boilerplate.
It's not a SaaS or a black box. Just clean, object-oriented Python code, with a built-in DRY_RUN mode to test strategies on live data without risking real money. Grab it, tweak it, do whatever you want with it.
👉 Get the full source code here
Disclaimer: This code is for educational purposes. Always test in
DRY_RUNfirst, protect your API wallets, and remember this is sold "as-is" with no tech support.
