App Pricing Research

app-pricing-research

First step of app pricing-strategy research: get an app's prices across many countries (original local price + estimated USD), one table per pricing tier.

Quick start

cd app-business-skills/skills/app-pricing-research/scripts

# default: ChatGPT across large-population markets
python3 benchmark.py

# a specific app + countries (ISO codes or names), write to a file too
python3 benchmark.py --app-id 6448311069 --countries in,br,id,tr,us --out report.md

# resolve an app by name (lists matches if ambiguous; re-run with --app-id)
python3 benchmark.py --app-name "Duolingo"

How to choose inputs

App — give --app-id if known. Otherwise --app-name "" resolves it via the iTunes Search API (if several match, it prints candidates — pick one and pass --app-id). No app specified → defaults to ChatGPT (6448311069).

Countries — --countries takes ISO-3166 alpha-2 codes (in,br,jp) or country names (India,Brazil,Poland). Any country resolves automatically from the bundled data/countries.json (~246 countries → name, flag, currency). An unknown name errors clearly; an unknown code is still tried best-effort. With no --countries, it uses the default_markets list in data/countries.json — edit that list to change your standing set of interested markets. To target the user's own top markets, first pull installs/proceeds by territory from App Store Connect (Sales & Trends needs a Sales/Finance/Admin key + vendor number; the rich Analytics Reports API needs an Admin key).

How it works

• Prices: the App Store listing is rendered per storefront with the gstack browse binary

(real JS render — IAP prices are NOT in the static HTML), then read from the Information → In-App Purchases section. URL form: apps.apple.com//app/id?l=en.

• USD: live rates from open.er-api.com (verified to track Google Finance within ~0.2%);

the report stamps the rate date. The currency is read from the displayed price (so USD-billed storefronts like Kenya convert correctly), falling back to the country's currency.

• Output: one markdown table per pricing tier (rows = countries; local price + ≈USD).

Notes / limits

• USD is approximate — FX updates ~daily and drifts between Apple's price-point updates.

The local price is exact (straight from Apple); treat ≈USD as a comparison aid.

• Duplicate tiers: some apps list the same IAP name twice (e.g. ChatGPT's monthly + annual

"Plus"); the tool disambiguates by ascending price as Name (1) / Name (2).

• Locale formats are handled (Rp 75ribu=75,000, 132.000đ=132,000, R$ 39,90=39.90).

If a price can't be parsed safely, ≈USD shows ? rather than a wrong number.

• Unknown / no data: an unrecognized country name errors; a country that returns no IAPs is

listed in a "No data returned" note at the end rather than silently dropped.

• Each country is a separate browser load (~a few seconds); ~15 countries takes a couple minutes.

Files

• scripts/benchmark.py — the tool. data/countries.json — country reference (code → name +

currency) and the editable default_markets list; regenerate from mledoze/countries to refresh.

Prereqs

• gstack browse (the /browse skill) installed.
• python3 (standard library only).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: vicoa-ai
• Source: vicoa-ai/app-business-skills
• License: MIT

Install and usage instructions live in the source repository linked above.