Tokens vs Cloud credits (how usage is counted)
What it is
AISA counts AI usage in one of two ways, depending on how you access AI. With your own provider keys (Free and BYO-key tiers) usage is measured in tokens — you pay your AI provider directly, in cents. In Cloud you use AISA's server and usage is measured in credits: 1 credit for a short request (one optimized content, one image) and 5 for a long generation (a Q&A set, a content rewrite, the llms.txt, a Consultant article). Every UI element — cost estimates, end-of-run reports, the progress and balance bars — follows whichever applies to you. The deciding factor is the access mode: credits (Cloud) or own (Free / Premium BYO-key).
Where to find it
AI SEO & GEO → Credits / AI Access to choose and view the mode. In Cloud, the credit bar is shown at the top of the Dashboard and across optimization screens; below 25% a top-up reminder appears. Cost estimates appear before any 1-Click / Bulk run.
Requirements & tier
Token counting applies to Free and BYO-key Premium tiers (Solo / Pro / Agency / Studio) and requires your own provider API key. Credit counting applies to Cloud, which includes credits and needs no key; top-up packs are available (e.g. €10/20/50 for roughly 4,000 / 8,500 / 22,000 credits) — the same rule applies to topped-up credits.
What a credit costs, by type of request
| Cost | What it covers |
|---|---|
| 1 credit | A short request: One-Click on one item, one field of the panel, an image alt text or title, one field fixed by the SEO Doctor, a summary, one rotation variant. |
| 5 credits | A long generation, the ones that produce a lot of text: a Q&A/FAQ set, Fix Content (rewriting the paragraphs), the llms.txt document, a SEO Consultant article. One-time setup steps — “Generate all” of the Brand Context and “Learn my style” of the Writing DNA — cost 1 credit even though they produce a long text. |
The cost is always shown before you click, right under the button. Inside a type the length of the text does not change the cost: a 300-word page and a 3,000-word page both cost 1 credit.
How to use
- Open Credits / AI Access and confirm your mode: Cloud (credits) or your own keys (tokens).
- Before any 1-Click or Bulk run, read the estimate — it shows credits (Cloud) or tokens (your keys) up front.
- In Cloud, keep an eye on the credit bar; under 25% you'll see the recharge prompt. Top up when needed — an in-progress queue resumes by itself, losing nothing.
- With your own keys, monitor spend on your provider's dashboard; One-Click batch (Pro/Cloud) cuts token use ~60% by doing everything in one call.
- To save credits in Cloud, turn off Auto SEO on publish so every publish doesn't trigger a paid call.
Questions people ask
- What is one credit? 1 credit = 1 short request (one AI call generates all of a page's SEO at once). The long generations cost 5.
- Why did a Q&A set cost 5 credits? Because it is one of the long generations: the AI writes several complete question/answer pairs in a single call. Fix Content, the llms.txt and a Consultant article are in the same group. The button says “(5 credits)” before you press it.
- Does a long page cost more than a short one? No. Inside a type the cost is fixed: One-Click on a 3,000-word page costs 1 credit exactly like on a 300-word one. Only the TYPE of request changes the price, never the length.
- What happens if I don’t have enough credits? Nothing is written and nothing is charged: the plugin tells you what the request costs and what is left — “This request costs 5 credits, you have 3” — so you can top up or run a short request instead. Those 3 credits are still perfectly usable for short requests.
- Can credit costs change? Yes. The number of credits per request type may change over time, for example when the prices of the underlying AI providers or market conditions change. The cost is always shown before you click, changes apply only to requests made after they take effect, and credits already in your balance are never reduced. See our Terms & Conditions.
- How much does it cost with my own keys? Cents for hundreds of items — DeepSeek is ~€0.0002 per generation; Mistral Small is similarly cheap.
- Why does One-Click cost less in tokens? It packs the whole page's SEO into one JSON call instead of three, saving roughly 60% of tokens; this batch mode is included in Pro/Agency/Studio and Cloud.
- If I top up mid-run, do I lose progress? No — recharge while a queue is running and it resumes automatically.
- Do my keys or content leave my site? Keys stay in your site's database. In Cloud, content passes through the AISA server only for generation.
- Why does each publish use credits? Auto SEO on publish makes one AI call per publish; disable it in Settings → Post Type if you don't want that.
Troubleshooting
UI shows credits but I use my own keys (or vice versa) → the access mode is wrong; switch it in Credits / AI Access.
Credits draining unexpectedly → Auto SEO on publish is on; disable it in Settings → Post Type.
Estimate looks high before a run → reduce selected fields/types, or use One-Click batch (Pro/Cloud) to cut token use.
Provider rate-limit (429) with your own keys → lower AISA Turbo parallelism or switch to DeepSeek for large jobs.
Related
Plans and tiers · Activate your license and get updates · Quick start: your first optimization