Kerith

Subprocessors

Last updated: October 20, 2026

This page lists the third-party service providers ("subprocessors") that process personal data on Kerith's behalf to deliver the Service. It is referenced by our Privacy Policy. We update this list when we add or remove a subprocessor.

Subprocessor Purpose Data categories processed Region
Cloudflare Core hosting and infrastructure: database (D1), object storage (R2: sermon audio, backups, exports), cache (KV), vector search (Vectorize), on-device transcription (Workers AI), edge rate limiting; AI Gateway, which routes requests to AI providers and keeps a short-lived log of request metadata (model, token counts, cost, a pseudonymous user identifier) without the content Account data, Your Content, push tokens, IP (transient) Global edge / US
Anthropic AI processing (chat, reflections, summaries) when a Claude model is used Content submitted to AI features US
OpenAI AI processing when a GPT model is used; audio transcription (Whisper) Content submitted to AI features; sermon audio US
Google (Gemini) AI processing when a Gemini model is used; sermon photo text extraction (Vision) Content submitted to AI features; sermon photos US
Groq AI processing when a Llama/Mixtral/Gemma model is used Content submitted to AI features US
Voyage AI Text embeddings for semantic search (RAG) Portions of Your Content US
Backblaze (B2) Cold-archive storage of encrypted weekly database backups Encrypted backup (all backed-up tables) US
RevenueCat Subscription management Subscription status and identifiers (no card data) US
Google Firebase Cloud Messaging Push notification delivery Push token, notification content US
PostHog Product analytics Usage events; account properties (denomination, experience level, subscription status, guest/registered, time zone) United States
Sentry (production only) Error monitoring Error metadata (content scrubbed) United States
Apple Sign in with Apple; App Store payments/subscriptions Authentication identifier; payment handled by Apple US

Note on AI providers. Which AI provider receives a given request depends on the model configured for that feature. Across all AI features, content is only sent after you grant AI consent in the app (see Privacy Policy Section 3.3).