aso-workbench README

项目的 README,来自 github.com/mustafakendiguzel/aso-workbench。合并为一个 markdown 文件:README.md。

ASO Workbench icon

ASO Workbench

App Store Optimization on your Mac, with honest numbers. Track where your app ranks, find keywords you can actually win, write and publish your listing in every language, and see what each change did.

Download macOS 15+ Swift 6.2 License

Install · Quick start · Features · Keys · How the numbers work · FAQ

English · Türkçe · 简体中文 · 日本語 · 한국어 · Español · Deutsch

ASO Workbench in action: tracking keywords, checking popularity and difficulty, writing a listing
▶ Watch the full demo · the preview above is silent and loops

Why ASO Workbench

Most ASO tools show one confident number per keyword. Often that number is a guess, and nobody tells you which part is Apple's and which part is the tool's.

ASO Workbench takes a different path:

What it does

📈 Track Rank for every keyword, per country and device, from real App Store search (top 200). History chart with your releases on it. Δ against your last release, so gains don't vanish when you re-measure. Ratings per country and written reviews.
🔎 Research Popularity from Apple wherever Apple publishes it, a labelled estimate elsewhere. Difficulty from the top ten results, fully explained. Golden keywords (real demand, beatable competition, not someone else's brand, relevant to your app). Import a competitor's keywords. Autocomplete ideas.
🌍 Go global Try your keywords in other countries in one pass, including AI-suggested phrasings in the local language, and track the winners with one click.
✍️ Write An AI writer that fills title, subtitle and keyword field with the keywords you can win first. Make better fixes every rule finding. Write every language at once, What's New for all languages, Terms and Privacy links added automatically.
🚀 Publish Send drafts to App Store Connect: missing languages are created, screenshots copied, the next version created, every write read back. Nothing is submitted for review.
🧪 Measure A timeline of releases and changes with a before/after report weighted by keyword value. Your Apple Ads results per keyword and country. MRR and subscriptions from RevenueCat.

Install

Download (easiest)

  1. Download ASO-Workbench-x.y.z.dmg from the latest release.
  2. Open it and drag ASO Workbench onto Applications.
  3. The first time, macOS says it can't verify the developer (the app is free and not notarized). Right-click the app → Open → Open. Or: System Settings → Privacy & Security → Open Anyway.

When a new version comes out, the app tells you (bottom of the sidebar, or ASO Workbench ▸ Check for Updates…). Your data is backed up automatically before a new version opens it.

Build from source

Needs macOS 15+ and Xcode 26 (Swift 6.2). No dependencies.

git clone https://github.com/mustafakendiguzel/aso-workbench.git
cd aso-workbench
make install        # optimized build → /Applications/ASO Workbench.app

Later: make update pulls the latest code and reinstalls. make help lists everything.

Quick start

  1. Add App (bottom of the sidebar) → paste your App Store link. The category is filled in for you.
  2. Keywords → Add → type the keywords you care about, one per line.
  3. Update ▸ Measure all → ranks arrive keyword by keyword (about 3 seconds each; Apple limits public search).
  4. Optional: Connections → add the Apple Ads key → Update ▸ Refresh popularity. Now you see Apple's own popularity.
  5. Filters ▸ Golden shows the keywords worth fighting for. Metadata ▸ Write turns them into a listing.

Connections: which keys, and why

You don't need any key to start. Each one unlocks something specific; the Connections screen explains each in the app, with step-by-step instructions.

Connection Need Unlocks Cost
(none) not needed Ranks, autocomplete, competitors, difficulty, ratings, reviews free
Apple Ads API Recommended Apple's own popularity, your campaign results free, spends nothing
App Store Connect API Optional Pull and publish your listing in every language free
OpenRouter Optional AI writer, translations, brand / relevance checks, popularity estimates pay per use, fractions of a cent
RevenueCat Optional MRR and subscriptions on Overview free
Apple Ads website session Advanced Apple's exact index for terms missing from the lists (undocumented, at your own risk) free

Keys live in one file on your Mac, readable only by you, and each key goes only to its own service. Details and how to get each key: docs/CONNECTIONS.md.

How the numbers work

The numbers, explained simply

Number Plain meaning Good is
Rank Where your app appears when someone searches this word. low (#1 is best)
Popularity (5–100) How many people search this word, on Apple's scale. high
Difficulty (0–100) How hard it is to reach page one, measured by how big and focused the apps already there are. low
Opportunity Popularity, discounted by difficulty: popularity × (1 − (difficulty ÷ 100)²). high
Fit How likely someone searching this wants an app like yours. related (≥ 50%)
Golden ★ Popularity ≥ 30, difficulty ≤ 45, not a brand, related to your app. ★

Popularity: Apple first, estimate second

Apple publishes a 1–100 popularity only for the top of each genre (monthly lists), and since 29 September 2025 reports 5 for everything below. ASO Workbench shows, in order:

  1. Apple's list value (badge Apple, needs the Apple Ads key),
  2. Apple's exact index (badge Apple · index, optional website session),
  3. a model estimate (badge Derived · est.), and for Apple's "5" a 6–29 rank among floor terms so you can still compare them,
  4. otherwise "< N": below the lowest listed score.

The estimate comes from a model we trained ourselves and ship inside the app:

  1. Data. About 55,800 search terms from five countries (US, GB, DE, FR, TR).
  2. Answer key. Each term is labelled with Apple's real popularity.
  3. Clues. Each term is described with public clues only: autocomplete, search results, and Jev. Jev (TypeSafe Jev, a small classification model reached through OpenRouter) answers narrow questions with probabilities, such as "is this a brand?", "does it sound like real demand?" and "which genre is it?".
  4. Training. Two small CreateML models learned to predict Apple's number from those clues. They ship with the app, so you never have to train anything.

In use, Jev answers the same questions for your new terms, which is why the estimate needs the OpenRouter key. On terms it never saw, the model ranks them in the same order as Apple with a correlation of 0.74 and is off by about 10 points on a typical term. It is an aid for ordering, always labelled, and Apple's number always wins. The full method and test results →

Difficulty

Computed from the top ten public results, the ones a searcher actually sees. It weighs how big the top three are (35%), how big the rest of page one is (20%), how many name the keyword in their title (20%), how established (10%), how actively updated (10%) and how well rated (5%) they are, with a discount when Apple fills the page with unrelated apps. The keyword inspector shows every part. Details →

Before / after

Every release (detected automatically), every change you publish and every note you add becomes a marker. The Changes page compares ranks before and after it, weighted by each keyword's value. Climbing 50 places on a term nobody searches counts for nothing.

Privacy and security

See SECURITY.md.

Project structure

Sources/
├── ASOCore/        pure logic: models, rules, difficulty, diffs, features (no network, no UI)
├── ASOProviders/   Apple, OpenRouter and RevenueCat adapters behind an allow-listed HTTP client
└── ASOWorkbench/   the SwiftUI app: screens, services, SwiftData model, design system
Resources/          Info.plist, icon, bundled popularity model
scripts/            bundle-app.sh (build the .app), release.sh (.dmg + GitHub release)
docs/               POPULARITY, CONNECTIONS, ARCHITECTURE, RELEASING

Three layers the compiler keeps apart, Swift 6 strict concurrency, zero dependencies. Architecture →

Development

make open      # debug build and launch
make test      # unit tests
make lab       # launch with the developer Lab (retrain the popularity model)
make dist      # build the .dmg and .zip for a release (maintainers)

Releases, versioning and how updates reach users: docs/RELEASING.md. Contributions: CONTRIBUTING.md. Changes: CHANGELOG.md.

FAQ

Why do my numbers differ from other ASO tools? Because Apple publishes very little below the top of each genre, other tools fill the gap with their own estimates, usually without saying so. ASO Workbench shows Apple's value when Apple has one and labels everything else.

Does connecting Apple Ads cost money? No. The popularity and report endpoints are read-only and need no running campaign.

Can it break my App Store listing? It writes only when you press Apply and confirm, refuses if the listing changed since you last pulled it, re-reads after writing, and never submits for review.

Why macOS 15? It uses SwiftData, the SwiftUI inspector and Swift 6 features that need macOS 15.

Where is my data? How do I start over? ~/Library/Application Support/ASO Workbench/. Quit the app and delete store.sqlite to start over; delete credentials.json to remove every key.

License

Free to use, not to sell. MIT License with the Commons Clause: you may use ASO Workbench for free (also for your own commercial apps), read and change the code, and share it with this license attached. You may not sell it, a modified copy, or a paid service built mostly on it. See LICENSE.

ASO Workbench is not affiliated with or endorsed by Apple. App Store, Apple Ads and App Store Connect are trademarks of Apple Inc.

English · Türkçe · 简体中文 · 日本語 · 한국어 · Español · Deutsch

aso-workbench (mustafakendiguzel/aso-workbench)

正在搜索数据库…

这是什么?ASO Workbench 是一款原生 macOS 应用,用于 App Store 优化(ASO)。它可以追踪你应用的关键词排名、挖掘有机会获得排名的关键词、用 AI 辅助撰写各语言的商店页面文案,还能直接把草稿发布到 App Store Connect。
这是网页仪表盘还是原生 Mac 应用?它是一款基于 Swift 和 SwiftUI 构建的独立原生 Mac 应用。所有功能都在你的电脑上运行,而不是在浏览器或远程服务器中运行。
我该如何让它运行起来?从 GitHub 最新发布版本下载 ASO-Workbench-x.y.z.dmg,打开后将 ASO Workbench 拖入“应用程序”文件夹,然后粘贴你的应用的 App Store 链接以添加应用,即可开始衡量关键词。
开始使用前我需要 API 密钥吗?不需要。关键词排名追踪、自动补全、竞品研究、难度评分和评论功能都无需任何密钥即可使用。可选密钥(Apple Ads、App Store Connect、OpenRouter、RevenueCat)可解锁真实的 Apple 热度数据、AI 写作与翻译、发布功能以及收入统计。
安装时有什么需要注意的坑吗?该应用未经过公证,因此 macOS 第一次打开时会拒绝运行。你需要右键点击应用并选择“打开”,或者前往“系统设置 → 隐私与安全性”,点击“仍要打开”来允许它运行。
我的 App Store Connect 和广告密钥究竟会发送到哪里?所有连接密钥都保存在你自己 Mac 上的同一个文件中,只有你本人可以读取,并且每个密钥仅发送给对应的服务。应用本身没有服务器,也不收集任何遥测数据。
什么样的关键词会被标记为“Golden”?当关键词的热度至少为 30、难度为 45 或更低、不属于他人品牌,并且与你的应用确实相关(即存在真实需求,而且你有机会胜出)时,该关键词就会被标记为“Golden”。
“Publish”会真的把我的更新提交给 App Store 审核吗?不会。“Publish”会把你的草稿发送到 App Store Connect,包括创建缺失的语言、复制截图以及设置下一个版本,但不会提交审核;提交审核的步骤仍需你自己手动触发。
我能为其他国家和语言研究并追踪关键词吗?可以。“Go global”功能可以让你一次性在多个国家测试你的关键词,包括 AI 建议的本地语言表达,并能一键追踪效果好的关键词。
为什么它不像其他工具那样给出一个综合 ASO 分数?作者有意将各数据来源分开:Apple 自身的热度数据、公开搜索测量结果和模型估算值各自都有独立标签,绝不会混合成一个猜测出来的单一数字。
大小
11.4 MB
星标
120
时长
1个月
提交 · 90 天