novquillAI usage policy
How novquill uses language models, what happens to your data, and what we deliberately avoid.
Models we use
For book analysis and character chat, novquill calls modern language models via Google Gemini (including Gemini 2.0 Flash) and OpenRouter as a fallback when another model or provider is needed. Character portraits may use a separate image provider (e.g. Flux via OpenRouter). The exact model depends on the task: speed, cost, and staying in the book’s voice.
Where data is stored
Account metadata, saved books, chats, and settings live in Supabase (PostgreSQL). Infrastructure is hosted in EU regions to align with GDPR expectations. Book text is processed to extract characters and run chat; we do not publish your files or expose them to other users unless you explicitly enable sharing (e.g. a public link you turn on).
How we protect data
Connections use HTTPS. Row Level Security (RLS) in the database limits each user to their own books and conversations. Passwords are not stored in novquill — Supabase Auth handles them. Data at rest is encrypted under our cloud provider’s standard infrastructure.
What we do not do
We do not use your uploaded books or chats to train public models. We do not sell conversation content to advertisers. Character replies are creative interpretation based on the text — not verbatim quotes and not a legally authoritative source.
Minors
novquill is not aimed at children under 16 without parental consent. If you believe a child created an account without permission, contact us and we will help delete the data.
Privacy contact
For privacy, account deletion, and personal data requests: hello@novquill.com. We will appoint a formal DPO as EU B2B scale requires; until then, requests are handled as data subject enquiries within reasonable time.