Service Providers and AI Processing
Last updated: September 7, 2026
This list supplements the Privacy Policy. It describes the current application integrations, not a certification of every vendor contract or a signed customer data-processing agreement. Not every provider receives every category or every request. Providers may use their own subprocessors; payment, sign-in and security processing may also involve independent controller roles.
Cloudflare — infrastructure, storage and delivery
Processes account and authentication records, documents, per-user chats/Style/Memory/usage, uploaded files, operational metadata, restricted backups and transactional email information. Services include Workers, D1, Durable Objects, R2 and Email Service. Processing uses distributed infrastructure; no particular storage country is promised here.
OpenAI — writing planning and assistance
Processes relevant prompts, conversation/document context, selected Style samples, Memory, supported attachments and derived text for requested writing and utility features, including planning the passage submitted to Humanize. Current calls use the Responses API with response storage disabled. API data is not used for training by default unless the API customer opts in. Abuse monitoring and feature-specific temporary retention can still apply. Blury does not claim zero data retention.
Weights & Biases / CoreWeave — rewriting and detection
W&B Inference hosts Blury’s rewriting and detector models. It receives text for rewriting, relevant request-derived instructions/context and text segments for detection, together with inference metadata. This can include personal information contained in the text.
Section 4 of W&B’s published AI Terms permits use of customer data to provide/support services, develop offerings and improve services, subject to applicable settings. Blury has not verified a separate account-specific no-training commitment or content-retention limit. This is a material difference from a blanket promise that no provider can reuse submitted content. Public vendor documents alone do not establish the terms of a particular signed arrangement.
Google — optional sign-in, Drive and support mail
Google sign-in supplies basic profile and authentication information. The optional Drive importer obtains an authorized file’s text; saved imports then follow Blury’s content processing. Google Workspace handles inbound support email and any attachments you send. Google authentication and consumer account services also operate under Google’s own privacy terms.
Google privacy policy · Google API Services User Data Policy
Stripe — payments and subscriptions
Processes account email, checkout and billing information, payment details, invoices, subscription changes, refunds and fraud signals. Blury receives limited billing identifiers and subscription/payment state. Stripe retains some records under its own legal, payment-network and fraud-prevention duties after a Blury account is deleted.