Privacy Policy
Last updated: September 2026
This Privacy Policy explains how BrainD ("we", "us") handles information in connection with our workplace AI platform (the "Service"). BrainD is designed to be self-hosted — for most deployments, your organization's content is processed inside your own infrastructure and is controlled by your organization's administrators.
Information we collect
- Account information — name, email address, and organization details you provide when you sign up or are invited.
- Authentication data — credentials and tokens used to sign you in (including via Google or other identity providers you choose).
- Connected content — data synced from sources you connect (documents, messages, tickets, etc.), together with their source permissions.
- Usage and diagnostic data — service logs, error reports, and operational metrics. This category covers how the Service itself is running; it does not include the contents of the documents, messages, or files you connect.
How we use information
- To provide, maintain, and secure the Service.
- To answer your queries with permission-aware retrieval over content you have connected.
- To authenticate users and prevent abuse.
- To communicate with you about your account and the Service.
We use usage and diagnostic data to operate, secure, and improve the Service. We do not use the contents of the sources you connect for any purpose other than delivering the features you use in the product.
Google user data
Where you choose to connect a Google account, BrainD requests only the access needed for the features you enable, and uses that data solely to provide those features to you inside the app.
What we request, and why
| Scope | Why the app needs it |
|---|---|
gmail.readonly | Read your mail so it can be indexed and returned as a cited answer when you search or ask a question. |
gmail.compose | Save a draft that you asked an assistant to compose, and send a message or a reply only at the moment you explicitly ask an assistant to send one. BrainD does not delete mail, and does not change labels or read state on its own. |
drive.readonly | Read the files you have access to, including Google Docs, Sheets, Slides and the structure of Google Forms, so their contents can be indexed, searched, previewed, and cited. |
drive.file | Work with the files BrainD itself creates for you, such as a folder you asked an assistant to create. It gives no access to your other files. |
calendar.readonly | Read events, your calendar list, and free/busy times so meetings and their context appear in search results and answers. |
calendar.events | Create, update or cancel an event, only when you ask an assistant to schedule or change one. |
documents | Create or edit a document, only when you ask an assistant to write or update one. |
spreadsheets | Create a spreadsheet or write values into one, only when you ask an assistant to. |
presentations | Create or edit a presentation, only when you ask an assistant to. |
forms.body | Create or update a form, only when you ask an assistant to. |
forms.responses.readonly | Read form responses so they can be searched and summarised. |
meetings.space.readonly | Read who attended the Google Meet meetings on your calendar, and their transcripts where one exists, so meetings can be indexed and cited. |
meetings.space.created | Create a meeting space, only when you ask an assistant to set up a meeting. |
Limited Use
BrainD's use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
Concretely, this means:
- Only to provide user-facing features. Google user data is used to deliver the search, retrieval, preview, citation, and assistant features that are visible to you in BrainD. It is not used for any purpose you cannot see in the product.
- Never to train AI models. We do not use Google user data — or any content from the sources you connect — to develop, train, fine-tune, or improve generalized artificial intelligence or machine learning models, whether ours or a third party's.
- Never for advertising. Google user data is not used to serve advertisements of any kind, including personalized, retargeted, or interest-based advertising.
- Never sold. We do not sell Google user data, and we do not transfer it to data brokers, information resellers, or any other party for their own purposes.
- No human reading. We do not allow humans to read Google user data, except: with your explicit consent for specific messages or files (for example, when you ask us to investigate a support issue); where necessary for security purposes such as investigating abuse; to comply with applicable law; or where the data has been aggregated and anonymized so that it no longer identifies you and is used only for internal operations.
Where Google data goes
In a self-hosted deployment, content synced from Google is stored in the databases and object storage that you operate, inside your own infrastructure. It is not sent to BrainD.
When you ask a question, the passages relevant to that question may be sent to the language model your deployment is configured to use, solely to generate the answer you requested. That model can be one you host yourself, one you point at with your own provider key, or the default model the platform provides. In every case the data is processed to answer your query and is not retained by us for model training. If you would rather no content leave your network at all, configure a self-hosted model.
AI and machine learning providers
Answering a question means sending the passages relevant to that question to a language model, so this section names every third-party AI service that can receive Google user data. The Limited Use requirements turn on exactly this, and a general assurance is not a substitute for naming them.
On the managed service we operate, requests are routed through OpenRouter (https://openrouter.ai/api/v1), a model gateway, to Google Gemini 2.5 Flash and Gemini 2.5 Flash-Lite for answering and summarisation, and to Meta Llama 3.1 8B Instruct for indexing. Embeddings are generated by OpenAI usingtext-embedding-3-large. Those are the only third-party AI services that receive connected content.
Training is excluded. Our accounts with these providers are configured so that content sent for processing is not used to train or fine-tune their models, and so that requests are not routed to providers which reserve the right to train on inputs. No Google user data, whether raw, aggregated, or derived, is used to create, train, or fine-tune any foundational or generalised machine learning model, whether ours or a third party's.
On a self-hosted deployment, the operator chooses the model. It can be a model you point at with your own provider key, or one you run entirely inside your own infrastructure. In that second case the content is processed locally and is never transmitted to a model provider at all.
Retention and revoking access
You can disconnect a Google connector at any time from the connectors screen, and you can revoke BrainD's access directly at myaccount.google.com/permissions. Disconnecting stops further syncing; content already indexed is removed according to your deployment's retention configuration, and your administrator can delete it immediately.
Data storage and security
In self-hosted deployments, your content lives in the databases and object storage you operate. We apply access controls, encryption in transit, and least-privilege principles. See our Trust Center for details on our security posture.
Sharing and disclosure
We do not sell personal data. Content is shared only with the subprocessors necessary to run the features you enable — for example, the language model or email provider your deployment is configured to use — and only to the extent needed to deliver that feature to you. Where the data in question came from a Google API, any such transfer is additionally bound by the Limited Use commitments above.
Data retention
We retain information for as long as your account is active or as needed to provide the Service, then delete or anonymize it in line with your deployment's configuration and applicable law.
Your rights
Depending on your location, you may have rights to access, correct, export, or delete your personal data. Because BrainD is self-hosted, most requests are fulfilled by your organization's administrators; contact them or us using the details below.
Changes to this policy
We may update this policy from time to time. Material changes will be reflected by updating the "Last updated" date above.
Contact
Questions about this policy? Email tech@braindd.xyz.