Use included hosted processing, connect your own AI provider keys, or run model inference through the Ollama companion on Hobby, Pro and Ultra. Compare provider terms, model quality and the complete data path for the documents you process.
Each successfully processed hosted page costs one PE. Free and Hobby provider-key runs use the monthly PE allowance; Pro and Ultra provider-key runs cost zero PE. Your provider bills your account directly with zero markup.
Local Ollama inference runs on your computer; account and workspace features remain hosted.