Subprocessors

Subprocessors

List of third-party service providers (subprocessors) used by SiteGPT to deliver our services.

Current Subprocessors

SiteGPT uses the following third-party service providers (subprocessors) to deliver our services. This list is maintained in accordance with our Data Processing Agreement obligations.

Last Updated: July 2026


Infrastructure & Hosting

Subprocessor Purpose Location
Cloudflare, Inc. Application hosting, content delivery, file storage, and security Global
Prisma Data, Inc. Database hosting (PostgreSQL) United States, Europe

Data Storage & Processing

Subprocessor Purpose Location
Convex, Inc. Backend database and application platform United States (AWS)
Pinecone Systems, Inc. Vector database for AI embeddings United States (AWS)
Upstash, Inc. Redis caching and queue processing United States (AWS)

AI & Machine Learning

Subprocessor Purpose Location
OpenAI, L.L.C. AI language model processing United States
Cohere Inc. AI embeddings and language processing United States

Content Processing

Subprocessor Purpose Location
Firecrawl Web content scraping and processing United States
Supadata YouTube content processing United States

Observability & Analytics

Subprocessor Purpose Location
Portkey AI AI observability and logging United States
PostHog Product analytics United States
DataFast Marketing analytics (cookieless) United States

Email & Communications

Subprocessor Purpose Location
Bento Email marketing and communications United States
AutoSend Transactional email delivery United States
Loops Transactional email delivery United States

Payments

Subprocessor Purpose Location
Paddle Payment processing (Merchant of Record; acts as an independent controller for payment data, not a subprocessor) United Kingdom

PHI Processing (HIPAA Deployments)

For customers with an executed Business Associate Agreement (see /legal/baa and /hipaa), PHI is scoped to end-user conversations (chatbot training content remains PHI-free by agreement), which restricts Protected Health Information to a minimal subprocessor chain, each member operating under its own BAA with SiteGPT before your BAA is executed:

Subprocessor Role in the PHI path
Convex, Inc. Application database (conversations, chatbot content)
OpenAI, L.L.C. AI responses and embeddings, zero-data-retention endpoints only
Pinecone Systems, Inc. Vector search over chatbot content

PHI does not flow to any other subprocessor on this page. These boundaries are enforced as preconditions of enablement: HIPAA processing is turned on for an account only once AI processing routes exclusively to OpenAI under zero data retention, conversation traffic is served directly from Convex infrastructure rather than through SiteGPT's general content delivery network, training content is confirmed PHI-free by agreement (which keeps content ingestion and queue subprocessors out of PHI scope), and notifications to observability, analytics, and email subprocessors carry no PHI (human handover notifications link agents to the dashboard instead of embedding conversation details). Until an account completes enablement, no PHI is permitted on it at all (see /hipaa).


Changes to Subprocessors

We provide at least 30 days' advance email notice to affected customers before adding or replacing a subprocessor that processes personal data on their behalf, and we update this page at the same time. Customers with a Data Processing Agreement in place may object as described in the DPA at /legal/dpa.

Questions

If you have questions about our subprocessors or data processing practices, please contact us at support@sitegpt.ai.