Freqtrade + OKX Setup Guide
A practical path from exchange API creation to a PAPER-tested Freqtrade deployment. The priority is controlled permissions and verifiable operation—not rushing to live capital.
1. Start with restricted API credentials
Create dedicated credentials for the bot. Grant only the permissions the deployment requires and keep withdrawals disabled. Store secrets outside source control and never expose the Freqtrade configuration publicly.
2. Configure the exchange and markets
Set OKX as the exchange, explicitly define the pair universe and confirm that trading mode, margin mode and pair naming match the instruments you intend to test. Spot and perpetual futures are different execution environments.
Permission model: treat the API key as production infrastructure
Spot and futures are different deployments
Do not treat a working spot configuration as proof that a futures configuration is correct. Contract naming, margin mode, leverage, liquidation risk, funding and exchange-specific order behavior add additional failure modes.
3. Run PAPER first
Enable dry-run and verify startup, market-data reception, signal generation, simulated entries and exits, fees and database persistence. A bot that starts successfully is not yet a validated strategy.
4. Validate operations
Restart recovery · API errors · rate limits · clock/timezone · order sizing · maximum open positions · stop behavior · logs · database backups.
Our implementation
Our V6 FreqAI experiment uses OKX market infrastructure in PAPER mode on BTC, ETH and SOL. Public metrics are exported separately so exchange credentials never reach this website.
View V6 PAPER results →Scope: Freqtrade + OKX integration
This page covers Freqtrade-specific use of OKX. For platform-neutral API permissions, credential isolation and exchange-side security principles, use our OKX API trading bot guide.