Subprocessors
The third parties that process customer data on StellarBase's behalf, what they do, and where they operate.
Effective as of 1 September 2026
What this page is
A subprocessor is a third party we engage to process customer data on your behalf when you use StellarBase. We keep this list public and current, and we announce additions or replacements in advance, as set out in our Data Processing Addendum.
For sensitive operational detail (for example, specific data-transfer mechanisms), we can provide further information on request under a non-disclosure agreement.
Infrastructure and hosting
| Subprocessor | Purpose | Data processed | Region |
|---|---|---|---|
| Scaleway S.A.S. | Cloud infrastructure, storage, and EU-hosted AI inference, including document summaries and descriptions of images inside documents | Customer data (databases, files) and derived content | France (EU) |
| ZONER a.s. | Hosting for our self-hosted AI models | Customer content in transit for model processing | Czech Republic (EU) |
| Google (Google Cloud, Gemini, Drive) | AI inference (Gemini); user-authorized file import from your own Google Drive | Customer content (prompts, documents); files you choose to import | EU / United States |
AI model and data-processing providers
You choose which model runs each request, so content you submit (including content imported from connected sources) may be processed by the provider you select.
| Subprocessor | Purpose | Data processed | Region |
|---|---|---|---|
| Anthropic, PBC | Large language model inference | Prompts and documents | United States |
| OpenAI, L.L.C. | Large language model inference | Prompts and documents | United States |
| xAI Corp. | Large language model inference | Prompts and documents | United States |
| Mistral AI SAS | Language model inference and document OCR | Prompts, documents, images | France (EU) |
| TensorX | EU-hosted language model inference | Prompts, documents, images | EU |
| Infercom | EU-hosted language model inference | Prompts and documents | EU |
| Black Forest Labs | Image generation and editing | Prompts and images, including an image you ask to have edited | Germany (EU) |
| Tavily, Inc. | Web-search tool available to agents | Search queries, which may include content you provide | United States |
Communications and billing
| Subprocessor | Purpose | Data processed | Region |
|---|---|---|---|
| Postmark (ActiveCampaign, LLC) | Transactional and service email | Name and email address | United States |
| Polar | Billing and subscription payments | Billing details and contact data | EU / United States |
Where we process your content
Several processing steps run on our own infrastructure rather than an external provider: text extraction (OCR), entity recognition, document embeddings for search, and anonymization. Other steps, such as image captioning and some model inference, use the external providers listed above; which providers apply depends on the model you select and the file type.
Tooling providers (no customer data)
The following providers support our operations but do not process customer data: GitLab (source code and CI/CD), Twingate (internal network access), and Hugging Face (one-time download of model weights).
Changes to this list
When we add or replace a subprocessor, we announce it at least 30 days in advance and update this page. You may raise a justified objection within 7 days of the notification, as set out in our Data Processing Addendum; if we cannot resolve it, you may terminate the affected Services. See also our Privacy Policy.