Customizing the Agent
After choosing a template or starting from scratch, you'll be guided through the customization workflow:
1. Agent Identity
Name & Icon: Choose a unique name and avatar.
Description: Add a public-facing summary if listing on the marketplace.
Tags: Use tags for discoverability (e.g., “scalping”, “AI/ML”, “DEX-only”).
2. Trigger Conditions
Time-Based: Schedule hourly, daily, or interval triggers.
Market-Based: Define specific price, volume, RSI, or volatility thresholds.
Event-Based: On-chain events like wallet movement, token listing, or whale buys.
3. Trading Logic
Action Chains: Define steps such as buy, sell, rebalance, stop-loss, or compound.
Risk Management: Configure stop loss %, trailing stop, take profit levels, and max slippage.
Fallback Behavior: Define how the agent reacts to unexpected input or failures.
4. Model Integration
Select from:
Straket Curated Models: Pre-trained, backtested models.
Upload Custom Model: Support for ONNX, TensorFlow Lite, PyTorch Mobile.
Fine-tune:
Learning rate (adaptive models only)
Model temperature (for LLM-supported agents)
Sensitivity & output confidence thresholds
5. Wallets & Connectivity
Wallet Whitelisting: Only interact with user-approved addresses.
DEX/Protocol API: Connect to supported DEXs (Uniswap, Raydium, Jupiter).
On-Chain Verification: Store trigger conditions and performance logs on-chain if desired.
Last updated