ASO Analyzer

Automated App Store Optimization audits for Google Play apps. Drop in your app name, your competitor's, and a few screenshots — get a prioritized action plan in your inbox in under a minute.

🔗 Try it live →

---

What it does

ASO consultants charge thousands for what is, frankly, a checklist run against your store listing plus some screenshot critique. This tool automates the 80% of that work that's mechanical:

• Scrapes your Google Play listing for title, description, ratings, and screenshots
• Compares against a competitor of your choice in the same category
• Scores screenshots with vision AI across hook strength, visual hierarchy, feature communication, and brand consistency
• Generates a prioritized audit with quick wins (this week), medium-term moves (this month), and strategic plays (this quarter)
• Emails the full report plus shows it in the browser

The output is a real, opinionated audit — not a generic ASO checklist.

How it works

[ Web form ] → [ n8n webhook ] → [ Jina (scrape) ]
                      ↓
                 [ GPT-4o-mini ] ← analyzes title/description/keywords
                      ↓
              [ GPT-4o-mini Vision ] ← scores screenshots
                      ↓
            [ Synthesizes audit report ]
                      ↓
       [ Google Sheets log ]   [ Gmail send ]   [ HTTP response ]

The whole flow is a single n8n workflow: form input → multipart upload → branching logic for "did we find the app?" / "do screenshots match?" → analysis pipeline → response. Errors at each branch return structured JSON with codes the frontend handles gracefully.

Tech stack

• Backend orchestration: n8n (cloud, free tier)
• Scraping: Jina Reader for Google Play listings
• LLM analysis: OpenAI gpt-4o-mini (text + vision)
• Storage: Google Sheets for audit logs
• Email: Gmail node
• Frontend: Vanilla HTML/JS + Tailwind via CDN, deployed on GitHub Pages
• API: multipart/form-data webhook, structured JSON responses with error codes

Repo structure

.
├── index.html              # The web frontend (single file, no build step)
├── workflows/
│   └── aso-analyzer.json   # Exported n8n workflow definition
├── docs/
│   └── screenshot.png      # README screenshot
└── README.md

Running it yourself

The frontend is just a static HTML file — host anywhere or open locally.

The backend requires:

1. An n8n instance (cloud or self-hosted)
2. Imported workflow from workflows/aso-analyzer.json
3. Credentials configured for: OpenAI, Google Sheets, Gmail
4. Active webhook URL pasted into the WEBHOOK_URL constant in index.html

API contract

POST /webhook/aso-analyze-webhook Content-Type: multipart/form-data

Fields:

Field	Type	Notes
appName	text	Your Google Play app name
category	text	e.g. "Education", "Photography"
competitor	text	Competitor app name
email	text	Where to send the report
Your_App_Screenshots	file (multi)	2+ images
Competitor_App_Screenshots	file (multi)	2+ images

Success response (200):

{
  "status": "success",
  "appName": "Marcus Chen Photography",
  "competitor": "ProCam",
  "email": "marcus@example.com",
  "audit_report": "...",
  "your_app_visual": "{...JSON-encoded scores...}",
  "competitor_visual": "{...JSON-encoded scores...}",
  "email_sent": true,
  "email_id": "..."
}

Error responses:

• 404 with code: "app_not_found" — couldn't locate the app on Google Play
• 422 with code: "screenshot_mismatch" — uploaded screenshots don't match the named app

Known limitations

• Reviews data is shallow. Jina scrapes the listing page but doesn't capture user reviews. Replacing this with a proper reviews API is on the roadmap.
• Keyword volumes are AI-guessed, not measured. GPT pattern-matches "this looks like a high-volume keyword" rather than pulling real search volume data. Needs an AppFollow / App Annie integration.
• Brand bias in vision scoring. When the model sees a known brand, it tends to score more generously. The fix (already prototyped) is anonymized "App A vs App B" prompting.

What's next

• Reviews scraper (separate API)
• Real keyword research integration
• Anonymized brand prompts in vision scoring
• Error sheet logging + metrics dashboard
• Prompt versioning for analysis quality tracking

License

MIT