Custom AI Assistants: Frequently Asked Questions
faq
Answers to the questions people ask before building a custom AI assistant — sources, privacy, accuracy, teams, and what 'no-code' actually means.
Before you commit to a custom AI assistant, you probably have a short list of the same questions everyone asks. Here they are.
What is a custom AI assistant?
An AI chat interface grounded in your sources — pages, PDFs, notes, screenshots — instead of the open web. It answers in your domain, not in general.
Do I need to write prompts or code?
No. In SNAP you save sources, create a project, and chat. There is no prompt template to tune.
How does it avoid hallucinations?
Every answer is generated from your snaps and cites them. If a source doesn't support a claim, you see that immediately. Read more on SNAP's source layer.
Who can see my sources?
Only you, by default. Sharing is opt-in — privately with a person, with a team project, or across your company.
What file types can I feed it?
Web pages, page regions, PDFs, screenshots, and voice notes. Anything you can capture in a browser.
Can a team share one assistant?
Yes — scoped Project workspaces let a team pool sources without leaking them elsewhere.
How fast is it to set up?
Under a minute for a first agent. Deeper coverage improves as you SNAP more sources.
More detail in the full custom AI assistant FAQ.