Need help? Check our guides or contact support.
AI Token Usage
AI features are the primary driver of token consumption on ikigize. This page explains exactly which AI operations cost tokens, how the price is calculated, and what remains free.
What is free
Not all AI-powered features cost tokens. These operations run without any token charge:
- Semantic search — finding courses, resources, and people by meaning
- Matchmaking — connecting learners with peers and mentors
- Content recommendations — surfacing relevant courses and resources
- Basic chat — standard messaging and collaboration
What consumes tokens
Token-consuming AI features are those that call external LLM providers or paid APIs on your behalf:
AI agent conversations Asking Iki, the Professor, or the Librarian for help. Cost depends on the length of the conversation and how complex the response is.
Content generation Using AI to draft module outlines, task instructions, learning objectives, or course descriptions. Longer, more structured outputs cost more.
Librarian cataloging When the Librarian agent maps a course, module, or resource to the Learning Graph — running topic searches, exploration, and relationship commits via a streaming pipeline.
Web search The Librarian and other agents can run external web searches to find public resources. Each search call incurs a small token cost.
Embeddings Creating vector embeddings for semantic search and matching. Typically a minor, one-time cost when content is first indexed.
The pricing formula
All AI costs follow the same formula:
Token Cost = (Variable Cost × 1.9) ÷ 0.01€The variable cost is the inferred cost from the underlying service provider (e.g. the LLM API). The 1.9 multiplier is the 90% markup for platform sustainability. Dividing by 0.01€ converts the euro amount into tokens.
Each token has a baseline value of 0.01€, providing predictable pricing and easy budget planning across all premium operations.
For all premium operations (AI interactions, storage), ikigize adds a 90% markup to the inferred cost, ensuring platform sustainability while maintaining transparency.
The token price remains fixed at 0.01€ per token. As underlying service costs fluctuate, token consumption rates adjust rather than token prices, maintaining financial clarity.
Clear cost breakdown for every operation. Users can see exactly how tokens map to platform costs, ensuring trust and informed decision-making.
Infer Variable Cost
Calculate actual cost from service providers
AI API costs, storage infrastructure costs
Apply 90% Markup
Add markup for platform sustainability
Variable cost × 1.9
Convert to Tokens
Convert final cost to tokens
Cost with markup ÷ 0.01€ per token
Token Cost = (Variable Cost × 1.9) ÷ 0.01€
Token Value: 0.01€ per token | Markup: 90%
Premium Operation Costs
Average AI Conversation
Variable: 0.10€ → With markup: 0.19€
19 tokens
Storage per GB/month
Variable: 0.021€ → With markup: 0.04€
~4 tokens
The token price is always 0.01€. If underlying AI costs change, the number of tokens consumed adjusts — not the token's euro value. Your budget planning stays predictable.
Practical estimates
| Operation | Approx. tokens | Approx. cost |
|---|---|---|
| Average AI conversation | ~19 tokens | ~0.19€ |
| Short content generation task | ~10–30 tokens | ~0.10–0.30€ |
| Librarian full catalog run | ~30–80 tokens | ~0.30–0.80€ |
| Single web search | ~3–5 tokens | ~0.03–0.05€ |
These are estimates — actual costs vary with conversation length, content volume, and model selection.
Which wallet is charged for AI
AI usage is always debited from your user wallet — the account that triggered the request. There are no separate course or campus wallets for AI.
If you are a member of an organisation that provides coverage for AI spend, part of the cost may be paid from the organisation wallet up to the limits your admins configure. Your wallet view shows your usage and how coverage applies. Coverage does not change the fact that usage is attributed to you; it only affects how the cost is split between wallets.
Token refunds
If an AI operation fails due to a platform or provider error, tokens are automatically refunded. The transaction history clearly marks refunded operations so your balance is always accurate.
Next steps
- Wallets & billing — user vs organisation wallets and how AI fits in
- Storage Billing — how file storage is billed separately
- Allocation & Limits — organisation allocations and member limits