If you use the built-in Rix AI, credits are the meter that keeps usage simple and predictable. You do not juggle model prices, tokens or provider invoices: one AI action is one credit, and your plan decides how many you get each month. This page explains exactly what costs a credit, what does not, how the special cases are priced and where to watch the number as you work.
What one credit buys
The base rule is deliberately plain:
- 1 credit = 1 AI action. Rewriting a paragraph, summarizing a section, asking the workbook a question, generating a chart description, a chat answer in the PDF editor — each action is one credit.
Two flows are metered differently because they cost the service differently:
| Action | Cost |
|---|---|
| Standard AI action (edit, rewrite, summarize, chat, analysis) | 1 credit |
| PDF-to-Word cloud conversion | 5 credits |
| AI deck pages within your plan's included allowance | Included |
| AI deck pages beyond the included allowance | 1 credit per 3 pages |
| RixSheets AI pages | Unlimited on all plans |
The PDF-to-Word conversion is a full cloud reconstruction of a document layout — fonts, tables, columns, images — which is why it is priced at 5 credits rather than 1. It is a batch job, not a chat action, and it is the only standard conversion that runs as a dedicated cloud process.
Deck pages, explained
AI deck generation in RixSlides is metered by page, not by request. Each plan includes an allowance of AI-generated deck pages per month. Building a deck inside that allowance costs nothing beyond your subscription. When you generate beyond it, the overflow is metered at 1 credit per 3 pages, so a twelve-page deck past the allowance costs 4 credits, not 12.
The per-three-pages rate matters for heavy presenters: it means the marginal cost of generating more slides stays gentle. Iterating on the same deck's pages counts as you go, so the allowance is best thought of as a monthly pool of generated pages rather than a limit on how many times you may press the button. If you generate a ten-page deck and regenerate three of its pages, you have consumed thirteen allowance pages in total.
Sheets are the opposite by design: RixSheets AI usage is unlimited pages on every plan. Whether your workbook has one sheet or fifty, page count never enters the meter. Analysis, formulas and chat in RixSheets consume actions like anywhere else, but the size of the file never scales the cost.
What never costs a credit
Credits only apply to the built-in Rix AI. The other two ways of running AI are always unmetered:
- Bring your own API keys. OpenRouter, Anthropic, Gemini, DeepSeek, OpenAI or a custom endpoint — usage goes directly to your provider under your own billing, and Rixcart never meters it. See BYO API keys.
- Ollama local models. Everything runs on your own machine, so there is nothing to meter at all. See Ollama local models.
If you find yourself using AI heavily every day, a BYO-key plan is usually the cheapest route, because provider pricing ends up below per-action credits at volume. The crossover point is different for everyone, but the usage history on the credit page tells you quickly which side of it you are on. Watch one full month of real work: if your credits run out before the month does, the arithmetic of switching has effectively been done for you.
How entitlements stay in sync
Your plan entitlements and credit balance are synced through a combination of Firestore and Cloudflare D1. Practically, that means the number you see in the desktop app matches the number on the website within moments, and a credit spent in RixDocs on your laptop is reflected immediately if you check from another device. Sync also covers plan features: deck page allowances, provider access and plan-gated AI features all resolve from the same entitlement record, so there is one source of truth rather than per-app counters that drift apart.
For you this removes a whole class of small annoyances. There is no per-device credit pool to reconcile, no stale balance after switching machines, and no case where the app and the website disagree about what your plan includes.
Where to see your usage
- Open the credit usage page on the web and sign in.
- Review credits used, credits remaining and the breakdown of what consumed them for the current period.
- In the desktop app, open Settings and check the plan and credits section for the same live balance.
- If you are consistently near the limit, compare plans in plans and pricing explained or move heavy workloads to a BYO-key or local setup.
How plans map to credits
The subscription plans differ mainly in monthly credit allowances and provider access:
| Plan | Monthly credits | Notes |
|---|---|---|
| Starter | 20 credits | Light AI use, built-in Rix AI |
| Standard AI | 99 credits | Regular writing and analysis |
| Pro API & Local | 399 credits | Adds BYO keys and Ollama local models |
| Premium API | 1,200 credits | Heavy AI workloads with BYO keys |
Each plan has its own documentation page — start with Starter, Standard AI, Pro API & Local or Premium API. If your month is predictable — a fixed set of reports, a weekly deck — match the plan to the count of actions rather than guessing upward.
What to explore next
- Plans and pricing explained — the full comparison of allowances and features.
- Built-in Rix AI — what your credits are actually paying for.
- BYO API keys — the unmetered alternative for heavy use.