“Token-based pricing” is becoming the default for AI products. It’s also the most-confusing pricing model most customers have ever seen. This article demystifies it for AIOX specifically: what a token is, how AIOX charges, why we picked this model, and how to budget.
What a token actually is
A token is the unit of work an LLM processes. Loosely: ¾ of an English word. The sentence “AIOX makes your content AI-discoverable” is about 8 tokens. A 1,000-word blog post is roughly 1,300 tokens. A short prompt + LLM response uses both input tokens (what you asked) and output tokens (what came back) — AIOX bills the sum.
Tokens vary by language. English is the most token-efficient. CJK languages and code use more tokens per character. Most AIOX customers operate in English so the rule of thumb holds.
Why AIOX uses tokens (not per-feature pricing)
Three reasons:
- It maps to our actual cost. Every AIOX app calls one or more LLM providers (Gemini primarily, with optional OpenAI / Anthropic). We pay per token. Charging per token aligns our pricing with our cost so we don’t have to subsidise heavy users or overcharge light ones.
- It’s portable across apps. One token pool serves AIOX Processing, AI Forensics, AI Visibility Score, AIOX Audit, Schema Generator — every app. You don’t have to predict which apps you’ll use most.
- It’s fair under variable workloads. A 200-word post costs ~300 tokens to process. A 5,000-word post costs ~7,000 tokens. Flat per-post pricing would either overcharge short content or undercharge long content.
How each app uses tokens
Approximate, real-world per-operation costs:
- AIOX Processing: 1,200-1,800 tokens per post (varies with length).
- AIOX Audit: 1,500-3,000 tokens per URL.
- AI Forensics: 8,000-20,000 tokens per investigation (4 providers × 10-15 probes).
- AI Visibility Score: 5,000-15,000 tokens per daily refresh (depends on topic count).
- Schema Generator: folded into Processing cost.
- Bot Sentinel auto-classification: ~50 tokens per unknown bot.
- Help Concierge questions: 200-1,000 tokens per Q&A, but not charged to you — the operator absorbs help chat costs.
What your plan covers
- Free: 10,000 tokens/month. About 7 post processes, or 1 Forensics investigation, or 1 daily Visibility Score for two weeks.
- Pro ($29/mo): 500,000 tokens/month. About 350 post processes, or 30 Forensics runs, or daily Visibility Score with room for everything else.
- Agency ($99/mo): 3,000,000 tokens/month. Multi-domain headroom — 25 domains’ worth of typical usage.
Token packs (one-time top-ups)
When your monthly allowance runs out (or you want to bulk-process an archive without changing plans), buy a pack:
- 100k tokens — $9 ($0.09 per 1k)
- 500k tokens — $39 ($0.078 per 1k)
- 2M tokens — $129 ($0.065 per 1k)
- 10M tokens — $499 ($0.05 per 1k)
Packs never expire. They’re used after your monthly plan tokens are exhausted.
How to budget
- Pre-app estimate is shown before every operation. If the budget would tip over, the app warns you with a token-pack CTA.
- The dashboard shows current burn rate and projected month-end usage.
- For bulk operations (e.g. archive processing), the estimate appears before the job starts.
- Tester accounts get unlimited tokens — useful for evaluating before committing.
