# AI that starts from your measurements URL: https://storelift.net/ai/ Language: en Updated: 2026-09-24 Storelift's AI does not start from a blank prompt. It reads what Storelift has already measured for your app — the keywords you track, where you rank on each, autocomplete demand, the apps above you — and writes or suggests from that. The model writes; the store's own numbers decide what is true. ## At a glance - 8 — AI features live today - 60 — AI visibility runs a month on Studio - 20 · 100 — localized listings a month on Pro · Studio - 8 — MCP tools, from Pro Every number on this page is read from the product's own code and plan limits, so it changes when the product does. ## AI visibility in Claude People ask assistants which app to use. Storelift asks Claude the questions a user would ask about your category — 6 per storefront, in up to 3 storefronts per app — and looks for your name in the answer. - Reads: the keywords you track, to write the questions, and the rivals measured above you, to know which names to look for. - You get: whether Claude names you, in what position, and who it names instead — dated, so the trend is the evidence, not a single answer. - Scored by search, not by the model: the answer text is searched for names; the model is never asked to grade itself. NOTE: Of the 19 other ASO and app-intelligence tools we checked on 24 September 2026, one measured AI visibility — as an enterprise add-on, in ChatGPT and Gemini. None measured Claude. Plan: Studio 60/month · In the app: AI visibility · How it works → ## Keyword suggestions from your own data The planner reads your app's category, its store description and the keywords you already track in a storefront, and suggests the terms a searcher would actually type there, in that storefront's language. It suggests; it does not judge. - Reads: your listing's genre and description, and your tracked keywords, so nothing is repeated. - You get: short candidate terms, without your brand name, ready to add to tracking. - The store decides: a term you add goes through the same measurement as every other — rank and demand come from the store, never from the model. Plan: Pro 100/month · Studio 600/month · In the app: AI keyword planner · How it works → ## Metadata drafts The Draft screen writes a title, subtitle, description and, on the App Store, promotional text for one storefront — using only terms measured there, the most-demanded first. Your brand name is kept. - Reads: your measured keywords with their current rank and demand, and your live listing as context. Terms that could not be measured are left out and listed. - You get: a draft counted against each field's character limit, with warnings for Apple product names in the title (App Store) and for pricing language. - Both stores: App Store and Google Play, each with its own fields and limits. Plan: Pro 100/month · Studio 600/month · In the app: Metadata studio › Draft · How it works → ## Localized listings: Global Localization Global Localization finds the App Store languages your listing does not cover and writes a complete listing for one of them: title, subtitle, the 100-character keyword field and description. Keywords are chosen from terms measured in that storefront and from Apple's own suggestions there, not invented. - Reads: your tracked keywords with measured demand, your source listing, your live App Store Connect keyword field when it is connected, and App Store autocomplete in the target storefront. - You get: a listing to copy field by field or download and, if you ask, your screenshot captions rewritten in that language. - Counting: one right is one language of one app. Plan: Free 1, once · Pro 20/month · Studio 100/month · In the app: Global Localization · How it works → ## Screenshot caption reading On the Screenshots screen, the model reads the marketing captions printed on your listing's first screenshots in a storefront. Interface labels, buttons and the status bar are left out. - Reads: the screenshot images the store serves in that storefront. - You get: each screenshot's caption as text, kept until the image set changes. - Not a ranking lever: Apple does not index screenshot text. Its value is conversion, and the screen says so. Plan: Pro 100/month · Studio 600/month · In the app: Store page monitor › Screenshots ## Review summaries and reply drafts Review AI reads your app's recent reviews in a storefront and tells you what users are saying: a short summary and the themes behind it, each with how many reviews it covers. Next to any review it can draft a reply in that review's language. - Quotes are real: every quote shown under a theme is checked word for word against the actual reviews; a quote that isn't there is dropped. - Counts are ours: a theme's number is the reviews matched to it, counted by Storelift, not a figure the model wrote. - Drafts, not posts: a reply draft is short, in the reviewer's language and promises nothing you haven't said. You copy it; nothing is sent for you. Plan: Pro 100/month · Studio 500/month · In the app: Reviews · How it works → ## Storelift AI: questions about your own data Ask in plain words — "why did my rank for this keyword drop in Germany?", "what do recent one-star reviews complain about?" — and Storelift AI answers by reading your tracked keywords, rank history, competitors, reviews and store page, the same data you see in the panel. - Numbers come from your data: the answer is built from what it read, and it says which data it read. - One answer, all your data: keywords, ranks, competitors, reviews and store page combined in a single answer. - Only your apps: it reads with your own account's access, so a limited seat sees only the apps it is allowed to see. Conversations stay in your browser. Plan: Free 5/month · Pro 50/month · Studio 300/month · In the app: Storelift AI ## Connect Claude to your data: the MCP server The MCP server lets Claude — or any MCP client — read your measurements directly: 8 tools for ranks, rivals, your store page, reviews, charts and AI visibility. Ask “which of my keywords dropped this week” and the answer comes from the same measurement the dashboard shows. - Claude.ai: add https://storelift.net/mcp as a custom connector and sign in — there is no key to copy. - Claude Code and other clients: one command, below, or the local storelift-mcp package with an API key. - Read-only: the connection reads the same data as an API key and cannot change anything. In Claude Code: claude mcp add --transport http storelift https://storelift.net/mcp NOTE: Several larger tools sell MCP access only with a separate API subscription. Here it comes with Pro, at $20 a month. Plan: Pro · Studio · How it works → · Setup guide → ## What the AI does not do The honest part ### The model writes. It never supplies a number. - It does not measure. Rank, demand and the apps above you come from the store's own search result and autocomplete. The model is never asked whether a keyword is good. - It does not publish. You copy or download drafts and localized listings; nothing is written to App Store Connect or Google Play. - It does not grade itself. AI visibility counts names in the answer text instead of asking the model where your app ranks. - It does not browse for the assistant. AI visibility measures the answer Claude gives from its own knowledge — the answer most people get. - It does not count an unmeasured term as a result. A suggested keyword is only a candidate until the nightly measurement gives it a rank and demand. - The MCP connection cannot change anything. It reads your data; it does not edit it. ## Which plan includes what | | Free | Pro | Studio | | --- | --- | --- | --- | | AI visibility | — | — | 60/month | | AI keyword suggestions | — | 100/month | 600/month | | Metadata drafts | — | 100/month | 600/month | | Global Localization | 1, once | 20/month | 100/month | | Screenshot caption reading | — | 100/month | 600/month | | Review AI | — | 100/month | 500/month | | Storelift AI | 5/month | 50/month | 300/month | | MCP server | — | ✓ | ✓ | | Price | $0 | $20/month | $60/month | Every AI feature runs on our key and is included in the plan. You do not need an API key of your own. ## FAQ Q: Which AI model does Storelift use? A: Claude, from Anthropic — for every feature here that calls a model. It runs on our key and is included in the plan; you do not bring your own API key. Q: Is any of the AI on the free plan? A: On Free: Global Localization and Storelift AI. From Pro: AI keyword suggestions, Metadata drafts, Screenshot caption reading, Review AI and MCP server. On Studio: AI visibility. Q: Does the AI change my listing in the store? A: No. You copy or download what it writes. Nothing is sent to App Store Connect or Google Play. Q: Can the AI make up search volume or rank? A: No. Rank and demand are read from the store: its search result and its autocomplete. A suggested term has no numbers until it has been measured. Q: Can I read my Storelift data in Claude? A: Yes, with the MCP server (Pro · Studio). In Claude.ai, add https://storelift.net/mcp as a custom connector; in Claude Code it is one command. The connection is read-only.