How Wsup AI is tested
The methodology behind every published claim about wsup.ai: which surfaces the platform is checked on, how the content toggle is exercised, how store metadata is verified against primary sources, and how often the coverage is refreshed.
Surfaces checked
Wsup AI runs across three surfaces, and behaviour differs across them, so coverage checks each one. Desktop web is checked in Chrome and Firefox from a European IP; the browser version is where the Filtered / Unfiltered toggle sits and where character creation runs. Mobile web is checked on iOS Safari and Android Chrome, with a specific note on the Add to Home Screen shortcut on iOS. iOS app is checked against the App Store listing at apps.apple.com/us/app/wsup-ai/id6753868146. Android app is checked against Google Play at package ai.wsup.app.android. Every claim that could shift across surfaces carries a surface tag on the overview page.
Regions checked
Primary testing runs from a European IP, because most of the audience for this coverage sits outside the US. Anything that touches the App Store or Google Play listings is cross-checked against the US storefront at least once, because store metadata (in-app purchase prices, publisher name, category tag) is region-locked and the US listing carries the authoritative pricing.
How the content filter is verified
The Filtered / Unfiltered toggle in chat settings is exercised in three ways. A soft ladder starts with PG-13 prompts and steps gradually toward explicit; the point at which the red warning banner appears is logged on both settings. A cold ladder resets the chat and drops explicit content on turn one, again on both settings. A drift ladder starts a Filtered chat and asks the character to shift tone gradually, to see whether the moderation follows the toggle or the content. Notes from all three runs feed the content policy section on the overview.
How pricing is verified
Credit-pack prices are read directly from the App Store and Google Play in-app purchase panels on the day of publication. Third-party review sites are not used as sources for prices, because they copy each other and stale numbers propagate for months. The credit tiers on the overview page ($1.99, $4.99, $9.99, $19.99) came off the US App Store listing on October 6, 2026.
How store metadata is verified
iOS metadata (App Store ID 6753868146, publisher, category, current version) is read directly from the App Store listing. Android metadata (package ai.wsup.app.android, publisher Skywork AI Pte. Ltd., last-updated date) is read from the Google Play listing. Install-count context comes from AppBrain and is treated as a secondary directional source, not a first-party figure; Google Play itself publishes install brackets rather than exact counts.
How the character builder is checked
Every field in Create Character is opened during coverage refresh cycles. The six personality sliders (Serious to Funny, Introverted to Extroverted, Formal to Casual, Stoic to Expressive, Traditional to Innovative, Transparent to Mysterious) are checked by building two characters with matching text prompts and opposite slider positions, then comparing the voice difference across a run of turns.
How WHOIS and domain data are checked
Domain registration data (June 16, 2024, Key-Systems GmbH) is read from public WHOIS records. Cross-checks come from independent scanner reports at Gridinsoft and Eveninsight, cited as secondary sources.
Source hierarchy
Primary sources: the Wsup AI interface at wsup.ai, the App Store listing, the Google Play listing, public WHOIS records, and content on the platform itself. Anything readable without paywalls or NDAs. Secondary sources: third-party review sites (TechLinos, Techraisal, AiBoxTools), community discussion threads, AppBrain install figures. Secondary sources are cited by name with a date, and never used as the sole support for a factual claim. Chain-of-review sites (reviews that cite other reviews without primary observation) are not used.
What is not tested
Latency and Cloudflare rate-limiting during US evening peak hours are not logged systematically from a European IP; the overview page's phrasing about peak-hour throttling reflects that limitation. Voice call quality is not tested, because Wsup AI does not offer voice as of July 2026. Long-term retention beyond a month is not tested; the memory section rests on session-length tests and secondary reports.
Refresh cadence
Full refresh every three months. Pricing and store metadata get a spot-check monthly, because both change fastest. Every published claim carries the date it was last verified on the affected page; the dateModified field in the JSON-LD updates whenever a claim changes. Structural rewrites (adding a section, removing a section) bump the review date too.
Where this methodology fits
This document sits alongside the rest of the site's editorial framework. Ownership, purpose and reader profile are covered on the about page; independence, affiliate handling and the corrections log sit on the editorial policy. Reader-side rules are on the terms of service, and data handling is covered separately in the privacy policy. The age verification policy explains the gate on entry, and the 18 U.S.C. § 2257 statement covers compliance for editorial text-only content.
Last reviewed: October 6, 2026 · by Melissa Blake