Billing.
Nilux uses a pay-per-use model. No subscriptions, no monthly fees, no hidden charges. You pay only for the tokens your requests consume.
How it works
- Top up your balance — add funds to your account
- Use Nilux — each request consumes tokens, measured in real time
- Cost is deducted — the exact token cost is subtracted from your balance after each request
- Never overspend — when your balance reaches $0, requests are blocked. You never go into debt
Pricing
Current rates, per-model breakdown, and an interactive cost calculator are on the pricing page. Prices are per 1 million tokens and include all infrastructure costs.
Smart caching
After the first request in a session, the system prompt and conversation context are cached. Subsequent requests reuse the cache — cached input tokens cost significantly less than fresh tokens. In typical coding sessions, the vast majority of input tokens are cached, keeping your actual costs well below the base input price.
Checking your balance
Use /balance inside a session to check your current balance and context usage. You can also view detailed usage statistics in the Dashboard.
When your balance runs out
New requests are blocked with a clear message. No debt is incurred. Top up your balance to resume.
FAQ
- How is token count measured? — Both input tokens (your conversation context) and output tokens (the model’s response) are counted using the provider’s tokenizer.
- What if a request fails? — You are only charged for tokens actually processed. Incomplete responses due to errors are not billed.