How Billing Works

Per-request billing with full transparency.

Per-request billing

Every API call is billed individually. After each request completes, we:

1. Measure the original and compressed token counts
2. Calculate the cost difference (savings)
3. Charge the actual LLM cost + 20% of savings (router mode) or just the 20% fee (BYOK)
4. Deduct from your prepaid balance
5. Log the full breakdown in your dashboard

Billing formula

original_cost = original_input_tokens × input_price + estimated_original_output × output_price
actual_cost   = compressed_input_tokens × input_price + actual_output_tokens × output_price
savings       = max(0, original_cost - actual_cost)
fee           = savings × 0.20

# Router mode
total_charged = actual_cost + fee

# BYOK mode
total_charged = fee

What you see in the dashboard

Each request in your usage history shows:

Field	Description
Model	The model used
Input tokens	Original → compressed
Output tokens	Actual tokens generated
Compression rate	% of input tokens saved
Savings	Dollar amount saved
Charged	Amount deducted from balance

Balance protection

• If your balance reaches $0, API requests return 402 Insufficient balance
• We never charge more than your available balance
• No surprise charges — you control how much you deposit

Monitoring usage

Track your spending in real time:

• Dashboard: opencompress.ai/dashboard shows balance, monthly stats, and per-request logs
• API: Usage data is included in the usage field of every response