Business judgment on AI products
On a webpage, summarize, rewrite, or translate without copy-paste or tab-switching — right-click and it's done.
01
Start inside the user's day · Public facts + observable behavior · 2026-08-28
On a webpage, summarize, rewrite, or translate without copy-paste or tab-switching — right-click and it's done.
Public materials do not yet show how users complete this job today or what they replace.
The product targets friction in this job, but public user evidence does not yet show the cost, frequency, or consequence of leaving it unsolved.
People already work in the browser; they will not open a chat window just to ask a model. The trend is assistants that live on the current page. The entry is the daily loop of reading, email, and foreign-language pages, sold as a monthly subscription.
A public record shows 2.56M monthly visits and 13311.5% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
① Whether next month's MoM holds — if 2.56M stays at 2M+ or rises, it was real growth; if it collapses back to the hundreds of thousands, the methodology change is confirmed; ② The overall browser-distribution layer — are peers (Merlin, Sider, Monica) moving in sync or trading share; ③ Any official …
Worth dissecting. A public record shows 2.56M monthly visits and 13311.5% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
compressing a high-frequency six-step copy-paste-switch routine into one gesture is the plainest efficiency logic there is. Before building a tool, count the steps users take in the scene it replaces; if you can remove two or more, the product has a reason to exist.
Freemium. Free tier gets daily-limited fast AI replies; the paid tier unlocks unlimited smart models and image generation. Price quotes are inconsistent across sources: the official pricing page shows roughly $12/month billed annually ($144… ① Whether next month's MoM holds — if 2.56M stays at 2M+ or rises, it was real growth; if it collapses back to the hundreds of thousands, the methodology change is confirmed; ② The overall browser-distribution layer — are peers (Merlin, Sider, Monica) moving in sync or trading share; ③ Any official …
On a webpage, summarize, rewrite, or translate without copy-paste or tab-switching — right-click and it's done.
A public record shows 2.56M monthly visits and 13311.5% month-over-month growth. That explains the attention, but product-level retention and payment are not yet verified.
An English validation note will follow from the public evidence.
The product claims to help users complete: “On a webpage, summarize, rewrite, or translate without copy-paste or tab-switching — right-click and”. User evidence has not yet verified pain intensity or the cost of doing without it.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
The assessment is recorded; an English explanation is pending.
02
Market comparison
The Chinese–English market comparison is not complete yet. A conclusion follows only after its coverage and verifiable evidence are recorded.
03
The call and next move come first; the full read retains the evidence and counterevidence.
An AI assistant pipeline for the browser: it stuffs ChatGPT, Claude, Gemini and more into a sidebar and right-click menu, so you can summarize, rewrite, translate, and ask without leaving the page.
A browser extension (Chrome/Edge) operated by MaxAI Inc, on the Chrome Web Store since 2023 and named to a16z's 2024 generative-AI consumer app ranking (Top 50). Founder, team size, and jurisdiction are not disclosed.
Read: it trains no models and owns no models — it is a distribution layer for models, plugging in whoever is best and charging for convenience. Its assets are install base and habit; its risk is that model makers build the same feature into their own products.
It positions itself as an in-UI assistant for a human, not an autonomous agent and not an API service — a deliberate boundary written into its own ToS.
The copy-paste-switch-tab routine.
To summarize a paragraph, polish an email, or translate a passage on a webpage, the standard sequence was: copy, open a new tab, paste into ChatGPT, wait, copy back, paste into the original spot. Six actions per operation, dozens of times a day. MaxAI compresses those six actions into one right-click or one click, keeping the workflow uninterrupted.
Read: it does not replace ChatGPT; it replaces the round trip to ChatGPT. That is the shared bet of every browser-extension AI tool: users are lazy enough about switching that they will pay for not switching.
Freemium. Free tier gets daily-limited fast AI replies; the paid tier unlocks unlimited smart models and image generation. Price quotes are inconsistent across sources: the official pricing page shows roughly $12/month billed annually ($144/year) or $30/month monthly; some reviews cite $19.99-20/month. Annual billing is non-refundable.
| Dimension | Call |
|---|---|
| Founder-product fit | Not disclosed; cannot be assessed |
| Product insight | "Uninterrupted workflow" is the right pitch, and multi-model choice gives users agency; but features are near-identical to rivals (Merlin, Sider, Monica) |
| Execution quality | The extension has run stably across major browsers for two-plus years with broad model coverage; adequate for a distribution layer |
| Timing | The plugin lane is crowded, and model vendors are shipping in-app assistants themselves, squeezing the middleman |
Worth watching — but that +13311% MoM needs a question mark. Backed out of 2.56M monthly visits, the prior month was roughly 19K. A jump of that size looks more like a ranking-methodology change (newly counted web app / new domain traffic, or a restat after a site redesign) than a product suddenly going viral; nothing in public sources matches — no major press, no campaign, no launch.
Read: between genuine takeoff (a platform feature, a market breaking out) and a measurement-methodology change, the odds favor the methodology change. Treat 2.56M as an unverified upper bound, not real growth to project from.
Strip away the MoM anomaly and the product itself is a mature, working browser distribution layer with a real base of 1M active users. Its core risk is not quality but position: once model vendors build the same functions directly into browsers and operating systems, how much room is left for the middle layer? It is also expanding onto its own web app, but there it competes with ChatGPT itself.
① Whether next month's MoM holds — if 2.56M stays at 2M+ or rises, it was real growth; if it collapses back to the hundreds of thousands, the methodology change is confirmed ② The overall browser-distribution layer — are peers (Merlin, Sider, Monica) moving in sync or trading share ③ Any official action proving reality (a release, press coverage, store-ranking change)
Product logic: compressing a high-frequency six-step copy-paste-switch routine into one gesture is the plainest efficiency logic there is. Before building a tool, count the steps users take in the scene it replaces; if you can remove two or more, the product has a reason to exist.
Multi-model aggregation positioning: refuse to take sides, hand users the choice of which model fits a task, and sell "you are using a top-model pool." For end users that selling point depreciates as every model becomes good enough.
Worth watching, but unproven. The product is a mature tool with a real installed base, but the 13,311% MoM is most likely an artifact of measurement, and the real growth signal has to wait for next month's data.
05
Use these searches when the official site is missing or the current link is only a lead.