How credits are calculated

Each (prompt × model) execution consumes credits in a strict $1 = 1 credit ratio.

Last updated: 2026-05-05

TL;DR — Pranas uses a strict $1 = 1 credit ratio. Each scan consumes credits based on the models and prompts you choose. The scan dialog shows an estimate before you launch. When credits run out, scan execution is blocked immediately — no grace period. Top up from Account → Billing.

Why credits exist (and why we're transparent about pricing)

Every AI Scan execution costs real money — Pranas pays OpenRouter, Google, Anthropic, OpenAI, and other providers for each model call. Credits are how we pass that cost through to you predictably and fairly. Two principles drive the credit system:

1. Strict $1 = 1 credit ratio. No hidden conversion. If your invoice says "$50", you have 50 credits. Period.
2. No grace period when you run out. This is intentional — surprise overage bills break trust. The system blocks new executions the moment your balance hits zero, prompts you to top up, and you decide if and when to spend more.

How credits are calculated per query

Credits per (prompt × model) execution depend on two factors:

1. Model tier — lighter models (Gemini Flash Lite, GPT-5 Nano) cost a fraction of premium models (GPT-5, Claude Opus, Gemini 3 Pro).
2. Prompt and response length — longer prompts and longer responses consume more credits.

The scan dialog always shows the estimated credit total before you launch, based on the prompts × models you've selected. Actual usage is shown per query in the scan detail after completion.

Indicative credit cost per query

Multiply per-query × number of executions in your scan to get the total.

Where credit info shows up

What gets billed and what doesn't

Auto-retries are never double-billed. The queue tracks attempts and bills only the consumed work.

When credits run out

The moment your balance hits zero:

1. The next queue worker that needs to dispatch a query checks your balance, sees zero, and refuses to dispatch.
2. Remaining queue rows that haven't started are marked as halted/blocked.
3. You see an upgrade/top-up prompt in the topbar and on the Run AI Scan dialog.
4. The dashboard still works fully — you just can't run new executions until you top up.

There is no overage billing. No surprise charges. You decide when to add more credits.

Plan-included credits

Each paid plan includes a monthly credit allowance:

Exact figures live on the pricing page and the Plans overview article — they may change. Your current balance is always the source of truth in Account → Billing.

Unused monthly credits do not roll over to the next billing cycle by default — use them or lose them. This is standard for usage-based platforms.

Top up

You can add credits any time without changing your plan:

1. Open Account → Billing.
2. Scroll to Credits.
3. Click Top up.
4. Choose an amount.
5. Complete checkout via Stripe.
6. Credits are added within seconds of payment confirmation.

Topped-up credits persist across billing cycles (unlike plan-included monthly credits). They roll over until used.

Worked example

Maya runs a 30 × 4 scan = 120 executions. The dialog estimates 18 credits ($18). Her balance is $25. She launches.

• 30 executions on Gemini Flash Lite: ~$0.5 total
• 30 on GPT-5 mini: ~$3
• 30 on Claude 3.7 Sonnet: ~$6
• 30 on Gemini 3 Pro: ~$8
• Final actual: $17.4 (slightly under the conservative estimate)

She finishes with $7.6 remaining. Next week she'll need to top up or rely on her monthly plan refill.

Edge cases & gotchas

• Estimates round up — actual cost is usually slightly less than the estimate.
• Premium models can 5–10× the cost vs. fast models. Always glance at the estimator before adding a Claude Opus or GPT-5 to your set.
• Long prompts cost more because input tokens scale linearly. Trim verbose prompts.
• "No brands found" still costs. Don't expect a refund for these — the model ran and produced output.
• Topped-up credits never expire, but plan-included monthly credits reset at cycle boundary.

Troubleshooting

Related articles

• What is an AI Scan?
• Choose AI models for a scan
• Scan statuses explained
• Plans overview
• Manage credits and balance
• Invoices and receipts

Still need help?

Email contact@pranas.co with your scan ID — we can break down exactly which queries cost what.