❓ Problem

AI collapsed the cost of shipping. Designers spin up clickable prototypes, marketers stand up landing pages, PMs build internal tools, and engineers ship side projects, all in an afternoon. The vibe-coding wave made this a flood, not a trickle: Lovable alone now mints over 100,000 new projects every day and crossed 25 million in its first year, and that is before you count v0, Bolt, Replit and Google AI Studio piling on millions more.

Here is the part that breaks. Every one of those builds lands in your Slack as a naked URL with no way to talk about it. So you do the same sad dance every time: open the link, take screenshots, number your feedback in a doc ("1. the header is off, 2. wrong shade of blue, 3. this button does nothing"), paste the lot into Slack, @ the owner, and then watch roughly none of it get actioned because nobody can tell which button "this button" actually was.

We solved this years ago for the work that lived in tools. Figma gave us comments pinned to the canvas. Google Docs gave us comments pinned to the paragraph. But the live web, which is now where more and more of the work actually ships, never got the memo. The artifact and the conversation about the artifact live in two different places, and the gap between them is where feedback goes to die.

✅ Solution

A browser extension that turns any website into multiplayer mode, so a team can comment and iterate on any web app, internal tool or prototype together.

• One click, comment on anything. Hit the extension on any page, highlight any element, and leave a threaded comment pinned to that exact spot. The owner replies inline and resolves it, and the whole thread lives on the page itself, not buried in a channel. Think Google Docs comments, except the "doc" is any URL on the internet.
• A built-in AI grounded in what you are looking at. Ask the assistant a question grounded in the page in front of you and your team's docs: "is this copy on-brand?", "what changed since the last review?", "does this flow match the spec?" You get an answer in context instead of tabbing away to go find it.
• Feedback that actually moves. Resolved comments route into Linear or Jira, ping the owner in Slack, and leave a clean audit trail of every review, so the loop closes instead of evaporating.

The wedge is narrow on purpose. Do not try to annotate the open web. Win the one place where there is a screaming, daily reason to comment: the flood of review-once AI artifacts that exist purely to get feedback and then get thrown away. The comment layer underneath expands from there.

📊 Key Numbers

Market size

• The "comment and collaborate on shared work" primitive is proven to carry billion-dollar businesses. Figma turned commenting on a canvas into around $1.06 billion of 2025 revenue and more than 13 million monthly users, two-thirds of whom are not designers. Adobe thought the category was worth a $20 billion acquisition offer back in 2023.
• The adjacent AI dev-and-build tooling slice that is generating all these artifacts is itself growing fast, estimated at roughly $4.86 billion in 2023 and forecast to reach around $26 billion by 2030.
• The current visual-feedback layer is real but sub-scale and fragmented: BugHerd, Marker.io, Pastel, MarkUp.io, Userback and Filestage, priced anywhere from about $7 per seat to $350 a month, almost all of it aimed narrowly at agencies and dev/QA teams. That is the floor this idea is trying to lift off, not the ceiling.

ARR potential

• Bottom-up: a per-seat team SaaS at roughly $15 per seat per month needs only ~250,000 paying seats to clear ~$45M ARR, and ~1,000,000 seats to clear ~$180M. With Lovable alone creating 100,000+ projects a day, the population of people shipping and reviewing AI artifacts is heading into the millions, so a seven-figure seat count over a few years is not a fantasy.
• Top-down: capturing even a low single-digit slice of collaboration and feedback tooling spend is a $100M+ outcome, and the comment primitive has a Figma-scale ceiling if it ever genuinely escapes the wedge.
• Realistic path: $10-30M ARR as a focused prototype-review tool, with a $100M+ ceiling if it becomes the default comment layer for AI builds. Treat anything beyond that as upside, not the plan.

⏰ Why Now