About NanoBook

NanoBook provides VRAM, Claude caching and RAG embedding cost calculators in English. We publish official sources, assumptions and verification limits.

Updated Nov 5, 2026

Calculate what it takes to run AI

NanoBook brings together tools for checking resources and costs before running AI. Current tools cover local LLM VRAM, Claude API 5-minute/1-hour caching and RAG chunking/embedding costs. Calculation assumptions and official sources are published alongside them.

The VRAM tool currently covers text generation fully loaded onto one NVIDIA GPU, with F16 KV cache and one concurrent request. Calculated estimates are not presented as measurements or speed guarantees.

Open VRAM calculator Compare Claude caching costs Calculate RAG embedding costs All tools

Distinguish checked data from unmeasured results

Model architecture and actual GGUF sizes are linked to their model cards, official libraries and sources.
Hardware measurements include the device, software versions, cache, context, placement conditions and observation method.
Unmeasured values are labeled as such. Model execution has not yet been measured on hardware.

Editorial standards

Operator information

Operator: kim hyun

Operating country: South Korea

Contact & corrections Advertising & affiliate disclosure