Pot să vibecodez AppGrowKit?

Ce construiești

Generate App Store screenshot panels from my real app screens with an image model, and poll Apple's public endpoints daily for my own apps' ranks, ratings, and reviews.

ce îți trebuie

• image model API key (fal.ai or OpenAI)
• vision model API key for the QA pass
• SQLite or Postgres
• a scheduler that actually runs daily

Submitted by the maintainer, who supplied the catalog figures from the production warehouse: 1.7M apps, 36 storefronts, 22M rating and 11M rank observations over 82 collection days, 65 GB on disk. Coverage is uneven by design, so the numbers are worth reading precisely: breadth is global, but deep multi-month rank history is concentrated in the apps and keywords the crawler prioritises rather than spread evenly across all 1.7M. Verdict written against the paid Pro tier; the DIY comparison targets a personal tracker for apps you own plus a screenshot generator.

Promptul

Un weekend cu un agent de coding. Golurile care rămân sunt mai jos, la „ce pierzi”.

Build me an App Store screenshot generator and a personal rank tracker to replace AppGrowKit, single user, my own apps only.
Use Node 22, TypeScript, SQLite (better-sqlite3), sharp, and a small React frontend on localhost. Do not offer alternative stacks.

- Screenshots: I upload my real app screens. Compose a 3-panel App Store set by sending each panel to an image model (fal.ai key in .env), passing my screens as reference images so the in-phone UI stays faithful.
- Send every render back to a vision model with a checklist: garbled text, elements sliced by a panel edge, invented star ratings or award badges, phone hardware where the style said none. If it fails, re-render once with the failures pasted in as corrections.
- Composite the headline as real text with sharp, not model-drawn glyphs. Resize output to exactly 1290x2796 (iPhone) and 2064x2752 (iPad).
- Tracking: a config file lists my App Store track IDs and the keywords I care about. A daily job hits Apple's public endpoints (itunes.apple.com/lookup, /search, the RSS chart feeds, and the customerreviews feed) with a real User-Agent and one request every 1.5 seconds.
- Store every poll as a row with a timestamp, never an upsert. Rank history is the whole point and you cannot recover a day you overwrote.
- The page shows my apps' rating and rank over time, new reviews since I last looked, and where I sit for each tracked keyword.
- Out of scope: competitor intelligence, keyword difficulty scores, and revenue estimates. Those need a catalog of hundreds of thousands of apps with months of history. Say so in the README rather than faking them from a single day of data.
- README: the two API keys, cost per screenshot set, and a warning that the tracker is worth little until it has been running for a month.

Ce pierzi

• months of rank, rating, and keyword history you cannot backfill
• the 1.7M-app, 36-storefront catalog behind competitor lookups and keyword difficulty
• revenue and download estimates, which are calibrated against that catalog
• keyword volume and competition scores, which need a corpus to be relative to
• the MCP server that answers ASO questions from Claude, Cursor, or ChatGPT
• an accuracy pass on generated screenshots: the QA critic, best-of-2 scoring, and exact-size output

De ce se plătește în continuare

Because the scrape is easy and the history is not. Anyone can pull today's chart positions from Apple's public endpoints; nobody can pull last quarter's. Scale compounds the same way: 1.7M apps across 36 storefronts is 65 GB of ClickHouse and a crawler that has been running for months, and a single-box copy polite enough not to get rate-limited spends a long time getting there. The paid tiers gate the catalog reads, not the AI, and that is the honest tell about where the cost sits. There is also real engineering in the screenshot pipeline that a one-sitting build skips: a vision critic that inspects each render for garbled text and sliced elements, two candidates scored against each other, and output resized to exact Apple dimensions. You can reach decent panels without that. Reaching consistent ones across a 6-panel set in three device formats is where the weekend goes.

moat: Date proprietare Scara infrastructurii ce e un moat

self-collected App Store time series; the history, not the scrape, is the moat

Cine l-a construit deja

Să pornești de aici e tot vibecoding: promptul e pentru când o vrei exact în felul tău.

• SerpBear (se deschide într-o filă nouă) — Self-hosted rank tracker for keyword sets you own. Web SEO rather than App Store, but the same daily-poll-and-chart shape. (alive)
• app-store-scraper (se deschide într-o filă nouă) — Node library for Apple's public lookup, search, chart, and review endpoints. This is the data layer a DIY tracker is built on. (alive)

Ești de acord?

Aplicații similare

AccuRanker 129 $ NU CHIAR Semrush 140 $ NU CHIAR ScreenshotOne 17 $ CAM AȘA