AI Notice

1. What AI-generated content is

Brainside uses third-party AI models — today mainly the Gemini family, through Google Vertex AI, and some models through OpenRouter — for chat, summaries, classification, reading documents and images, voice transcription, agents and semantic search.

AI-generated content is a suggestion. It is not professional advice of any kind.

2. Limitations

When using Brainside's AI, keep in mind that it can:

• Give answers that look right but contain errors;
• "Hallucinate", that is, make up information that is not in your data;
• Reproduce biases present in the model's training data;
• Misread the context of your question or your data;
• Be out of date about recent events.

3. Your responsibility

Review, validate and confirm AI-generated content before using it to decide or act. We are not liable for consequences of using it without that check.

4. What the AI saves to your memory

In AIs connected over MCP, Brainside instructs the AI to follow these rules when writing to your memory (Brain Node) — an external assistant may not follow them:

• New, durable knowledge — a decision, a fact worth keeping — it saves on its own and tells you; the same goes for adding text to a note or filling in an empty field;
• Before replacing or deleting anything that already exists, and before saving sensitive data (such as health, religion, identity documents or bank details), content from an external source (an email, a web page, or someone else's file or message) or data belonging to a client or employer, it asks first and saves only after you say yes;
• The list of what must not be saved only changes when you ask;
• It never saves passwords, tokens, API keys or other secrets, even if you ask.

You can review, edit and delete what was saved at any time.

5. Use of your data

Your data goes to the AI provider only to run the feature you asked for. We do not use your data to train AI models. Details in the Privacy Policy.

6. Third-party AIs connected over MCP

When you connect an external assistant (such as Claude, ChatGPT or Codex), the answers are generated by that provider's model, not by Brainside. Brainside delivers and writes data in your account at the assistant's request, with your authorization, and logs the writes in the audit trail.

The quality of the answers and the assistant's handling of data follow its provider's terms. Brainside does not control that processing, but it is responsible for securely delivering the data you authorized and for logging the writes. You can revoke access anytime (see the Privacy Policy, Section 5).

7. Ongoing improvements

We work to improve the quality and accuracy of our AI. Models and providers may change over time; this page will be updated when they do.

8. Contact

Questions about AI in Brainside: suporte@brainside.ai.