Can Jarvis Draft Replies To Design Review Comments In My Voice?

Yes. Jarvis (getjarvis.eu), the screen-aware desktop AI assistant for macOS, Windows, and Linux, drafts replies to design review comments that sound like you, then posts them back through its Figma connector once you approve. Press Cmd+/ (Ctrl+/ on Windows), point it at an open thread, and ask for a reply that accepts the change, pushes back politely, or asks a clarifying question. Routing to frontier models from Anthropic, OpenAI, or Google, it matches the tone you have used before because its persistent, user-controlled memory holds your phrasing and recurring reviewers. It can answer several stakeholder comments in one pass so a busy review day clears faster. Honest limit: Jarvis drafts and, with approval, posts text replies; it does not redesign the frame the comment is about, so a note like "tighten this spacing" still needs your hands in Figma. Free plan (40 requests/week) to try it, Pro $16/month.

Jarvis does not produce generic corporate filler. It learns the way you write from earlier drafts and from explicit instructions you give it, then keeps that in persistent, user-controlled memory. So when you ask it to reply to a stakeholder who wants a bolder headline, the draft reads like you on a normal day, not a template. You can steer per reply: warmer for a nervous client, firmer when you are defending a deliberate choice, terse for an internal teammate. It routes to frontier models from Anthropic, OpenAI, or Google to keep the wording natural. You review every draft before anything is posted back to Figma.

Review days pile up dozens of comments. From the floating bar you can ask Jarvis to draft replies to every open comment on a file at once, each tailored to what that specific note asks for. It can accept simple approvals, ask for missing detail where a request is vague, and politely decline changes that break the design system, citing your reasoning. With Slack connected it can also ping the reviewer that you have responded. Batching turns an hour of context-switching into a single approve-and-go queue, and memory means recurring reviewers get consistent, on-brand answers without you re-explaining your style each time.

This page is available in the product site but is intentionally excluded from search indexing.

Capabilities