Newtown Platform — Sub-processors
Last updated: 2026-10-13
This list forms part of the [Data Processing Addendum](./data-processing-addendum.md). We give customers at least 30 days' notice before adding or replacing a sub-processor, sent to the notice address on file and reflected here.
To join the notification list, email privacy@newtowndata.io. Notice to the address on file is given whether or not you subscribe.
Core infrastructure
| Sub-processor | Entity & location | What it does for us | Data it may process | |---|---|---|---| | Google Cloud Platform | Google LLC / Google Cloud EMEA Ltd. — United States (us-central1; provisioned BigQuery datasets default to the US multi-region) | Application hosting, batch jobs, the control-plane database, object storage for business context and compiled documentation, secret storage, and the BigQuery warehouses | All Customer Data; all account data | | Google (Identity) | Google LLC — United States | Sign-in for customers and staff, and the identity federated behind our OAuth 2.1 server | Workspace user email address and email domain | | Google Workspace | Google LLC — United States | Our own email, documents, and support correspondence, and the channel through which contractual notices are sent | Anything you send us or we send you, including support material that may contain Customer Data |
Data pipeline
| Sub-processor | Entity & location | What it does for us | Data it may process | |---|---|---|---| | Fivetran | Fivetran, Inc. — United States | Primary ingestion. Hosts the authorization page where you enter source credentials, holds the loader service-account key for your warehouse, and syncs source records into it | Your source system credentials (entered directly with Fivetran, never through us) and all records those connectors sync | | GitHub | GitHub, Inc. (Microsoft) — United States | Hosts your generated dbt project as a private repository in Newtown's organization, and runs the CI that lints generated SQL | Data model SQL, tests, documentation, schema and column names. Not table row contents |
AI
| Sub-processor | Entity & location | What it does for us | Data it may process | |---|---|---|---| | Anthropic | Anthropic PBC — United States | The model our data-modeling and surveying agents call, when a Newtown operator runs a pass and on a nightly schedule for every active workspace | Accumulated business context, warehouse schema, table and column names, metric definitions, dbt model source, change-request text, and the SQL text of a query asked through the customer's AI client. Sampled values from low-cardinality columns and example join-key values only where the customer enabled "Sample values when profiling" for that table (off by default). Bulk row-level extracts are not included |
Our agreement with Anthropic prohibits training on data we submit through its API. We do not train models on customer data ourselves. The query-serving path — the MCP endpoint that answers your questions — calls no language model at all.
Commercial
| Sub-processor | Entity & location | What it does for us | Data it may process | |---|---|---|---| | Stripe | Stripe, Inc. — United States | Payment processing and subscription billing | Billing contact name and email, payment method, billing address, transaction history. We never receive or store full card numbers |
Not sub-processors
Your AI client (Claude, ChatGPT, or another). We do not select it, do not contract with it, and cannot impose terms on it. When you connect one to the MCP endpoint, query results travel to that provider under your own account and your own agreement with them. You are the controller for that transfer. Read their retention and training terms before connecting a workspace that holds sensitive records.
Your data sources. Stripe, Shopify, QuickBooks, your payroll provider, your bank feed, and the rest are your vendors, not ours. Note that Stripe appears in two capacities: it is our payment processor, listed above as a sub-processor; and it is separately a data source many customers connect, in which capacity it is their vendor.
Your own warehouse. If you connect a warehouse you own — currently BigQuery or Snowflake — it is your infrastructure under your own agreement with that provider, in the location you choose. We access it with the credentials and roles you grant, and where you have selected Newtown-Managed Transformation we also write into the pre-agreed target schemas described in the Terms. It is not a Newtown sub-processor: we neither host it nor contract for it, and your agreement with that provider governs it.
What we do not use
Security questionnaires ask, so: we run no third-party analytics, product-analytics, session-recording, error-tracking, advertising, marketing-automation, CRM-enrichment, or customer-data-platform tooling against the application. There are no tracking cookies and no advertising pixels. The product sends no automated transactional email — workspace invitations are issued by a person, and notices are sent from our own Google Workspace accounts.
Questions: privacy@newtowndata.io Newtown Analytics LLC · New York