Embedding Cost Estimator
Size the cost of turning your corpus into vectors — the one-time pass and the ongoing re-embedding — against the €0-per-token local alternative.
Estimated cost
An estimate, not a quote. Excludes storage, re-ranking and retries; real cost depends on your models and pipeline. Cloud figures use the USD rate you set; the local option runs on your own infrastructure (no per-token charge). RAGSuite supports both — hosted providers and local Ollama embeddings.
Good to know
How is the corpus size estimated?
Documents × average words, converted to tokens at ~1.33 tokens per word. Embedding cost is corpus tokens × your rate. Re-embedding only the share that changes each month gives the ongoing figure.
Why is the local option €0 per token?
Open embedding models (e.g. via Ollama or a local server) run on your own hardware. There’s no per-token charge — you pay for compute you already own. That’s the line that doesn’t scale with corpus size.
Are the rates accurate?
They’re editable example list prices (June 2026). Set them to your provider’s current numbers. Nothing is fetched or stored.
Want this guaranteed in your own infrastructure?
Get a copy of these results by email and see RAGSuite running on a setup like yours — citation-backed, self-hosted, EU-ready.
Thanks — we’ll be in touch within one business day (DE/EN).
Embed once, or embed locally — your call.
RAGSuite supports every embedding route, including local models via Ollama, so a growing corpus doesn’t mean a growing per-token bill.