Overview
Budget limits are a critical safety feature that cap how much an API key can spend. This protects against bugs, unexpected behavior, and compromised keys.How Budget Limits Work
- Agent submits an order
- LuckyLobster calculates order cost:
price × size - Compares against remaining budget
- If sufficient, executes and updates usage
- If exceeded, rejects with budget details
Configuring Budgets
Configure budget limits for each agent at Dashboard > Manage Agents.Budget Settings
The effective budget is the most restrictive of all configured limits and the wallet balance.
Example Configurations
Budget Enforcement
Order Placement
When placing an order, budget is checked:What Counts Toward Budget
Checking Your Budget
Via Dashboard
View budget usage at Dashboard > Manage Agents for each agent.Via API
How Budget Is Calculated
Theusdc field is the actual spendable amount, determined by the most restrictive constraint:
limitedBy field tells you which constraint is active: "wallet", "fixed_limit", "percent", or "position_limit".
Position Limits
In addition to spending budgets, you can configure position limits:
These prevent concentration risk and limit exposure.
Best Practices
Start Conservative
Start Conservative
Begin with a low budget ($50-100) and increase as you verify agent behavior. It’s easier to increase limits than recover from losses.
Choose Appropriate Period
Choose Appropriate Period
- Use
dailyfor high-frequency strategies that need fresh limits each day - Use
monthlyfor steady strategies with predictable spending - Use
neverfor fixed-allocation strategies
Monitor Usage
Monitor Usage
Regularly check budget utilization in the dashboard. If an agent consistently hits limits, it might need adjustment or investigation.
Set Up Alerts
Set Up Alerts
Configure notifications (coming soon) for when budget usage exceeds thresholds like 80%.
Increasing Limits
To increase an API key’s budget:- Go to Dashboard > Manage Agents
- Click the budget settings for the agent
- Update the budget configuration
- Save changes
Dry Run Testing
UsedryRun: true in orders to test budget validation without spending: