Brainy
AI is off by default
A Brainy deployment ships with its AI provider set to None. In that mode, every AI feature returns a plain "no AI provider is configured" message and nothing you write is ever sent anywhere. AI only activates when an operator explicitly configures a provider (OpenAI or Azure OpenAI) and only runs when you personally click an AI action — it never runs in the background on your content.
Which providers can be configured
- OpenAI — the public OpenAI API (api.openai.com).
- Azure OpenAI Service — a private Azure OpenAI resource and deployment.
There is no other AI subprocessor. If a deployment is configured with a provider, every AI call goes to that one provider using one model, set by the deployment's configuration.
What is sent, action by action
Brainy currently offers five AI actions. Each one sends only the text relevant to that action — never your whole workspace:
- Summarize a note — sends that note's content.
- Suggest PARA placement & tags — sends that note's title and content.
- Extract action items — sends that note's content.
- Detect duplicate notes — sends the note you're checking (title and content) plus the titles and existing summaries of your other candidate notes.
- Generate an output draft (post, report, brief) — sends the output's title, type, and the full content of every source note you selected for that draft.
Each action is initiated by you, on demand, for a specific note, task, or output you are looking at — you always choose the scope by choosing which item(s) to run it on.
What comes back, and how it's labeled
Every AI result is returned with the model that produced it and a prompt version identifier, and Brainy keeps these attached to the generated content separately from your own edits. AI-generated suggestions, summaries, and drafts are shown labeled as AI-generated in the interface, alongside the sources that produced them, so you can see what came from you and what came from the model.
AI can be wrong — review before you rely on it
Language models can misread context, omit important details, or state something confidently that is not true. Brainy does not treat AI output as fact: it is a draft. Review AI-generated summaries, suggestions, and outputs against their sources before you act on them, share them, or let them replace your own judgment.
Training
Brainy does not use your content to train its own models, and does not fine-tune any model on your data. Whether a configured third-party provider (OpenAI or Azure OpenAI) itself retains or trains on API request content is governed by that provider's own API terms, which apply independently of this page — Azure OpenAI's data handling terms differ from the public OpenAI API's, which is one reason the choice of provider is an explicit deployment setting rather than fixed.
Limiting or avoiding AI
Every AI feature is optional and explicit — there is no AI action that runs without you choosing to trigger it. If you would rather nothing you write is ever sent to an AI provider, simply do not use the AI actions described above; the rest of Brainy (capture, PARA organization, tasks, Today, search) works fully without them.
See also the Privacy Policy for how the rest of your data is stored, and Account and data to export or delete your data at any time.