Spending Caps & Token Accounting
SemeSeam checks month-to-date recorded usage locally before each paid call and halts when configured thresholds are reached.
How SemeSeam Guards Your Budget
Deep reasoning runs and automated descents can generate hundreds of turns. Without protection, a run could quickly burn through expensive model credits. SemeSeam implements three layers of cost defense:
Calls to local Hermes 3 via Ollama are completely free and never consume tokens from your budget.
Before each paid call, SemeSeam checks your recorded usage for the current month against your cap. If the cap has been reached, the carrel halts further paid calls. (Because usage is recorded from provider tokens after each call completes, a single call may slightly push total usage past the cap).
Input and output token pricing is populated with editable default rates. You can adjust rates anytime in Settings → Spending to match your provider tier.
Configuring Your Caps in the Carrel
In the carrel desktop app, open Settings → Spending:
- Monthly Cap ($): Enter a dollar limit (e.g., $10.00). SemeSeam checks recorded month-to-date spending before each paid call and halts new calls once reached.
- Monthly Token Cap: Enter a raw token limit (e.g., 2,000,000 tokens) as an extra guard.
- Setting either value to
0disables that specific ceiling.