Last updated: October 27, 2026
Restoration Services Group LLC operates RubyRestore.ai. These are our principal third-party service providers that process customer data on our behalf to run the Service, grouped by what they do. We share with each provider only what it needs to perform its service, under contracts or terms of service that govern how it may use that data.
The third-party tools you choose to connect to the Service (for example your job-management, documentation, accounting, or email systems) are not on this list. They are independent services you control, and your use of them is governed by your agreement with each provider.
We add, replace and remove providers from time to time, AI model providers in particular, and keep this page current with our principal providers. Questions can go to privacy@rubyrestore.ai.
| Provider | Purpose | Data involved |
|---|---|---|
| Neon | Managed PostgreSQL database, hosted on Amazon Web Services in the United States | All Service data, stored and scoped by customer |
| Render | Application hosting, background processing, cache, document conversion, and application logs | Service data while it is being processed, and request logs, which can include user identifiers and the addresses of pages requested |
| Amazon Web Services | File storage and hosting for the app and website | Uploaded files, generated documents, and site assets |
Requests to an AI model can include your messages, playbook instructions, and relevant portions of your connected integration data. Which provider handles a request depends on the model tier chosen for a chat or set for a playbook.
Our Acceptable Use Policy requires that the AI features be used in accordance with the usage policy of the provider behind them. Each provider's policy is linked below.
| Provider | Purpose | Data involved | Usage policy |
|---|---|---|---|
| Anthropic | Claude models: the default for playbooks and scheduled work, and available in chat | Request content and model output | https://www.anthropic.com/legal/aup |
| OpenAI | GPT models: the default for chat. Also the text embeddings that let Ruby recall your saved preferences | Request content and model output; saved preference text | https://openai.com/policies/usage-policies |
| Meta | Muse models, available as a model tier | Request content and model output | https://ai.developer.meta.com/legal/acceptable-use-policy |
| Baseten | Hosted open models, available as model tiers. Also writes generated documents and reports, and conversation titles | Request content and model output | https://www.baseten.co/terms-and-conditions (and each hosted model's own license) |
| xAI | Grok models, available as a model tier | Request content and model output | https://x.ai/legal/acceptable-use-policy |
| Celeris | Models available as a model tier | Request content and model output | https://celeris.ai/terms (section 3) |
| OpenRouter | Backup routing to the same models when a primary provider is unavailable | Request content and model output, only while a backup route is in use | https://openrouter.ai/terms |
| Other AI model providers | Models we evaluate or use from time to time, including models reached through OpenRouter | Request content and model output | The provider's own usage policy |
| Provider | Purpose | Data involved |
|---|---|---|
| E2B | Isolated sandboxes where the assistant runs code and scripts | The files and data the assistant passes into the sandbox |
| Brave | Web search | Search queries |
| Provider | Purpose | Data involved |
|---|---|---|
| Composio | Authorization broker for Gmail, Outlook, Slack, Google Workspace, Microsoft 365, and other connected accounts | Authorization tokens and the data those connections return |
| Unipile | WhatsApp connection | Messages and attachments in connected WhatsApp chats |
| Browserbase | Secure browser sessions for the assistant's own login to systems that have no API | Login session data and the pages the assistant views |
| Telnyx | Receives verification-code text messages for the assistant's own logins | Text message contents |
| Twilio SendGrid | Transactional and notification email | Recipient names and addresses, and message content |
| Google Firebase Cloud Messaging | Push notifications in the mobile app (relayed to Apple devices through Apple's push service) | Device tokens and notification content |
| Capgo | Updates to the mobile app | Device and app version information |
| Provider | Purpose | Data involved |
|---|---|---|
| Stripe | Subscription billing and payment processing | Billing contact, plan, and transaction data; card details are handled directly by Stripe |
| Rollbar | Error tracking | Error reports, which can include the signed-in user's identifier and the request that failed |
| FullStory | Session replay inside the signed-in app | Recordings of how you interact with the app |
| Google Analytics | Website and app analytics | Page views, events, and device information |
| Slack | Delivers the in-app feedback you submit to our team | Feedback text and ratings, and the page you were on |
| HubSpot | Customer relationship management for our own sales and marketing | The name, email address and company you give us on our website or at signup, the website pages you visited before, and your account's plan and billing status |