TokelixWhat is Tokelix?
Tokelix brings configured LLM channels together behind one API. Use your API key with your tools, check input, output and cache prices in the model catalog, and inspect each request charge in the console.
Every key belongs to an organization and its creator. Organizations can prioritize lower prices, administrator-selected routes or provider price caps. Failover stays within routes that match the requested model and strategy.
Charges retain the selected channel, model, price version and token categories. Later price updates do not change completed request bills. Organizations keep separate balances, budgets and usage records.
Team administration
Manual credits record the amount, operator, note and unique business reference.
Administrators manage organizations, members manage their own keys and usage, and billing members review bills and statistics.
Organization and member budgets follow calendar months in Beijing time, with atomic reservations for concurrent calls.
Filter settled requests by date, member, key and model, with separate input, output and cache usage.
Choose a verified email to receive a reminder when the balance falls below your threshold.
How are charges calculated?
A price snapshot is saved at admission, and funds and monthly budgets are reserved before generation. Requests settle actual usage after completion; upstream failures refund the reservation. Active requests may settle slightly above a budget, while new requests stop once the limit is reached.
| Record | Tokelix |
|---|---|
| Input and output | Charged using model prices and actual token usage |
| Cache | Cache reads and writes recorded separately |
| Credits | Manual credits with idempotent business references |
| Routing | Budget, curated and custom price caps |
| Historical bills | Preserves the request price version after price updates |