Tools
LLM tools: cost, memory and context calculators
Each tool shows its formula, uses only published rates or numbers you enter, and refuses to give a total when a value it needs is missing. Nothing you type leaves your browser.
Token cost calculator
Cost per request, day and month from a published rate or your own, with cached tokens counted once.
API price comparison
One workload priced across every listed provider, with same-weights rows marked and CSV export.
VRAM estimator
Weights, KV cache and overhead, using each model's real attention layout from config.json.
Context window planner
Fit system prompt, history, documents and reply inside a provider's actual limit.
How the tools handle your data
The calculations run in your browser. Inputs are kept only in the address bar after the # sign, so you can bookmark or share a setup; that part of the address is never sent to our server. The tools store nothing else and set no cookies. Rates and model data are the same records shown in the price table and the catalogue, each with its source and check date.