Released 2026-06-04. A nicer AI chat experience, and a clearer reminder that AI output should be verified.
Changed
- AI replies now appear progressively (a typing effect) in the AI assistant, the playground AI chat, and agent chat — including the public embedded agent widget — instead of appearing all at once. Nothing about how the AI works changed; it just reads more naturally.
Added
- A short reminder under every AI chat: "AI can get things wrong — double-check important info," with a Learn more link to the AI Disclosure. It always shows on the embedded agent widget, even when the "Powered by FlowFn" footer is hidden.