_The TR trio (Claude, Perplexity, Gmail) is the most indie-actionable set this month: highest anger, lowest localization bar, weakest incumbent attention._
Claude by Anthropic (US) — 42 recent 1–2★ (42% of 100) · Top Grossing #2
- Users hurting on: credits burned on failed/unwanted output (6, med) — "It's programmed to burn through credits"; launch-crash loop on the July build (5, high); bans with no explanation (3, med).
- Asking for and NOT getting: account switching (2 requests), project-file viewing on iOS (1), read-only Cowork on mobile (1), working in-app search (1).
- Your opening: a fading #2 whose paying users publicly count wasted tokens — any AI-workspace tool that shows exactly what each request will cost before it runs wins their trust by default.
cross-checked: kept 3, cut 0
Dropbox: Cloud File Storage (US) — 38 recent 1–2★ (38% of 100) · Top Grossing #4
- Users hurting on: unauthorized charges and un-cancellable plans (6, med) — "Dropbox charged me around 180USD"; paywalls on viewing shared/own files (4, med); files deleted without consent (3, high).
- Asking for and NOT getting: in-app delete/empty-trash (1), granular/wildcard search (1), fix photo downloads failing (1).
- Your opening: a steady giant whose recent reviews are about billing trust, not features — simple honest storage/transfer with one cancel button undercuts it.
cross-checked: kept 3, cut 0 (single-user-plan claim removed; insufficient distinct evidence)
Speechify – Text to Speech (US) — 29 recent 1–2★ (58% of 50) · Top Grossing #7
- Users hurting on: the scan-a-book ad promise vs page-by-page reality (12, high) — "the ad is an outright lie, in order to listen to a book you have to scan every page"; $140/yr price revealed after a signup quiz (9, high); free voices that "sound terrible" (2, med).
- Asking for and NOT getting: diverse narrator voices (1), a monthly plan (implicit in 9 pricing complaints), continuous playlists (1).
- Your opening: 58% of recent reviews are 1–2★ at a steady #7 — a TTS reader with honest onboarding, a monthly tier, and decent free voices converts its refund-seekers directly.
cross-checked: kept 3, cut 0
Yahoo Mail - Organised Email (US) — 25 recent 1–2★ (50% of 50) · Top Grossing #9
- Users hurting on: post-update freezes/blank screens (5, high); spam that 3,500 blocks can't stop (4, high); vanished folders/contacts (5, high).
- Asking for and NOT getting: spam control without the $4.99/mo tier (1), deletions that stick (1).
- Your opening: a fading legacy inbox charging for spam relief — a lightweight mail client whose spam filter just works picks up 30-year veterans already announcing their exit.
cross-checked: kept 2, cut 0
ChatGPT (UK) — 54 recent 1–2★ (27% of 200) · Top Grossing #1
- Users hurting on: image generation dying midway (7, high) — "Starts creating the image, then stops, provides no information"; environmental backlash (7, high) — "I've switched to EcoGPT because this is one of the main ai systems that is destroying our planet"; ads served to paying subscribers (5, high) — "You either serve ads OR you charge. You don't do both"; shrinking free limits (5, high) — "Free version used to be great but the limit reset time is ridiculous now"; app broken/frozen (6, med).
- Asking for and NOT getting: ad-free experience for payers (1), more free image credits (2), consistent faces across edits (1).
- Your opening: still a wall (3.82★), but the ad rollout is the first crack, and the UK eco-boycott cluster is absent from US reviews — position against "ads in your AI" before the resentment matures.
cross-checked: kept 3, cut 1 (consolidated quality/moderation into app-broken cluster)
Claude by Anthropic (UK) — 68 recent 1–2★ (45% of 150) · Top Grossing #2
- Users hurting on: quality/instruction-following collapse (12, high) — "Claude is now unusable for simple work tasks. It ignores you, hallucinates, loses content easily"; broken app shell (9, high) — "Whenever I open the app it says something went wrong and doesn't load"; limits even on paid plans (8, high) — "paid for a subscription after three messages it tells me to upgrade"; guardrail paternalism (8, high) — "Never in my 49 years have I been told \"now shoo, go to bed\" at 8am in the morning!".
- Asking for and NOT getting: the newest model at parity with the US (3 region-gating complaints, med), voice/TTS that doesn't cut off (2), mobile project sync (1).
- Your opening: fading hardest of its three storefronts (0.37) while gating UK users behind ID checks and model delays — any assistant that treats the UK as a first-class region inherits its leavers.
cross-checked: kept 3, cut 1 (removed rolled-back Sonnet claim; insufficient evidence)
Grok - AI Assistant (UK) — 74 recent 1–2★ (37% of 200) · Top Grossing #3
- Users hurting on: weekly/paid rate limits (16, high) — "Barely generated 20 \"speed\" images and 8 quality images before being rate limited"; free tier gutted + upsell pushiness (12, high) — "for free users, prompts went from 30, to 20 and now 10"; Imagine quality regression (7, high) — "Why has the intelligence dropped so badly, it cant separate people from backgrounds"; UK moderation overreach (6, med) — "this app rejects all images of clothed people because UK laws".
- Asking for and NOT getting: character consistency in video (2), transparent limit meters (1), the daily-limits switch reversed (multiple).
- Your opening: a fading #3 that changed the deal after purchase at every tier — an image/video tool with published, fixed quotas takes its cancelling subscribers.
cross-checked: kept 3, cut 1 (removed voice-personas claim; marginal evidence)
Dropbox: Cloud File Storage (UK) — 121 recent 1–2★ (60% of 200) · Top Grossing #4
- Users hurting on: billing traps (15, high) — "I had £90 removed from my account for an upgrade I neither asked for or wanted"; sync/reliability regressions (13, high) — "it's become very unreliable syncing across devices"; storage upsell blocking access to own files (9, high) — "Won't let me look at my photos without tell me I've ran out of storage!!!"; UX complexity (8, high) — "It's turned into a corporate tool to sell".
- Asking for and NOT getting: a usable free tier (2), a 500GB/1TB mid plan (1), per-account colour themes (1).
- Your opening: 60% negative at a fading #4 — the single widest file-storage opening in this report; win on cancel-anytime billing and a mid-size plan Dropbox refuses to sell.
cross-checked: kept 3, cut 1 (removed unreachable-support claim; count below threshold; iPad-feature gap removed; insufficient distinct evidence)
Microsoft Word (UK) — 89 recent 1–2★ (45% of 200) · Top Grossing #6
- Users hurting on: subscription wall / entitlement failures (14, high) — "I've been paying for this for months and when I need to make a document it's asking [me to buy]"; keyboard-vanishing, crash and lag bugs (13, high) — "Update has made keyboard disappear when using Edit mode - fix ASAP please!"; iPad/iPhone UX inferior to desktop (10, med) — "It feels way harder to use than it should be"; data loss despite autosave (5, med); Copilot/AI intrusion (5, high) — "Word has hijacked my ability to control what I write by inserting words I have not chosen".
- Asking for and NOT getting: an off switch for Copilot/correction bubbles that actually works (1), an iPad UX that matches desktop (10 complaints).
- Your opening: a fading staple whose mobile editor is paywalled and buggy — a fast, reliable, one-price document editor for iPhone/iPad meets 89 reviewers' stated need.
cross-checked: kept 3, cut 0
Surfshark VPN (UK) — 27 recent 1–2★ (27% of 100) · Top Grossing #8
- Users hurting on: VPN IPs detected/blocked by websites (11, high) — "their VPN IP addresses are constantly misidentified by websites"; new UI removed features — favourites, double-hop, safe search (4, high) — "The option for double hop vpn has disappeared from the app".
- Asking for and NOT getting: restore favourites ordering (1), fix auto-connect issues (1).
- Your opening: surging (0.69) and mostly loved — not a target to attack, but its feature-stripping redesign shows even winners shed users when they delete power-user features; the opening is narrow and time-limited.
cross-checked: kept 2, cut 0
Perplexity - AI Search & Chat (UK) — 52 recent 1–2★ (52% of 100) · Top Grossing #9
- Users hurting on: credit system gutted Pro (14, high) — "Clicked on free trial, they took £20 & within an hour of playing I need to add credits"; accuracy/quality decline (9, high) — "It gets stuck in loops. It ignores what you say"; iPad app frozen for a week+ (8, high) — "iPad app has been frozen and unresponsive for nearly a week"; ads/upsell inside a paid product (3, med); phone-number wall on existing subscribers (3, med).
- Asking for and NOT getting: stop Computer popups (1), option not to auto-open Computer (1).
- Your opening: steady on chart but 52% negative — an answer-engine with flat, honest pricing and a working tablet app flips its Pro refugees.
cross-checked: kept 3, cut 1 (removed refund-refusals claim; insufficient distinct evidence; lost-history claim merged)
Microsoft OneDrive (UK) — 45 recent 1–2★ (90% of 50) · Top Grossing #10
- Users hurting on: file operations broken since update — move/multi-select/delete crash (12, high) — "Since the last update you can't move files on iOS"; MS Lens shutdown fallout — no document scanner replacement (5, high) — "Wanted to use MS Lens but I cannot scan anymore w/o this app"; sync/upload failures + corrupt files (5, high) — "photos videos just disappear, can't play videos can't save files to devices".
- Asking for and NOT getting: duplicate-photo finder (2), delete more than 20 images at a time, whole-folder download (1 multi-item plea).
- Your opening: 90% negative at a fading #10 — the Lens shutdown orphaned a scanning use-case worth targeting directly with a simple scan-to-PDF-to-cloud tool.
cross-checked: kept 3, cut 1 (removed notification/photos-app-nagging; Copilot intrusion; Account/Vault access claims; below confidence threshold)
Claude by Anthropic (TR) — 28 recent 1–2★ (56% of 50) · Top Grossing #2
- Users hurting on: app won't open / "Something went wrong" (5, high) — "Hata veriyor Something went wrong" (gives the error — same loop as US/GB); no Turkish UI (4, high) — "Uygulama arayüzünde dil seçemiyoruz büyük bir eksiklik" ("we can't pick a language in the app UI, a major gap"); limits/pricing (4, high) — "Free den pro geçtik free kadar soru soramadım hemen limit doldu" ("upgraded free→Pro, couldn't ask even as many questions as free before hitting the limit"); unexplained bans, appeals ignored (3, med) — "hiç bir sebeb yok iken ban yedim … 200 dolara claude pro almıştım" ("banned for no reason … I'd bought Claude Pro for $200"); quality — superficial/erroneous answers (4, high).
- Asking for and NOT getting: Turkish interface language (3 — the dominant request), Cowork on iPad (1).
- Your opening: a fading #2 grossing app in TR with no local language — an indie AI assistant that is Turkish-first (UI, prompts, support) attacks the exact gap its own reviewers spell out; this is the cleanest TR-only opening in the category.
cross-checked: kept 4, cut 1 (removed under-18-age-gate claim; only 2 supporting reviews, below threshold)
Perplexity - AI Search & Chat (TR) — 25 recent 1–2★ (50% of 50) · Top Grossing #9
- Users hurting on: credit system + price hikes (10, high) — "1 aylık pro paketi aldım 4000 kredi verdi,3 gün demeden kredi bitti" ("bought 1 month of Pro, got 4,000 credits, they ran out in under 3 days"); paid accounts suspended, refunds refused (3, high) — "Suspended paid access, refused refund or human review"; quality (5, med) — "Verdiğiniz paraya değmez" ("not worth the money you pay"); crash/freeze since last update (1, low).
- Asking for and NOT getting: cheaper regional tier (1 explicit request: "$5-7 a month"), account deletion support (1).
- Your opening: 50% negative and TR purchasing power makes the credit model radioactive — a lira-priced, flat-fee AI search tier wins this storefront on pricing alone.
cross-checked: kept 3, cut 1 (removed mandatory-phone-verification as standalone; news-notification-spam claim too marginal)
Gmail - Email by Google (TR) — 108 recent 1–2★ (72% of 150) · Top Grossing #10
- Users hurting on: account-switch freeze since last update (18, high) — "Son güncellemeden sonra hesaplar arasında geçiş yapılmıyor, donuyor uygulama" ("since the last update you can't switch accounts, the app freezes"); login/app won't open — endless spinner (13, high) — "Dönüp duruyor ama açmıyor" ("it just spins and never opens"); attachment/compose bugs incl. Turkish-encoding failure (9, high) — "ISO-8859-9 Türkçe karakter kodlamasıyla gelen ekler aç[ılmıyor]" ("attachments with Turkish ISO-8859-9 encoding won't open"); app bloat — 700+MB weekly updates (6, med) — "Neden Haftada 2 devasa güncelleme (700+mb) neden neden neden…?" ("why two giant 700+MB updates a week?"); storage-full paywall blocking mail even after paying (5, high) — "İki defa ücret ödediğim halde mail alamıyorum" ("I paid twice and still can't receive mail").
- Asking for and NOT getting: unread-only filter on All Inboxes (2), bulk delete (2) — "10 yıldır yorumlara sadece toplu silme yazı[lıyor]" ("for 10 years these reviews have asked only for bulk delete").
- Your opening: 72% negative at a fading default — a lightweight Turkish-first mail client (fast account switching, unread filter, bulk delete, correct TR encoding) answers a decade of ignored requests on this storefront.
cross-checked: kept 5, cut 1 (removed iOS-version-gating; only 4 supporting reviews; notifications-not-arriving merged to connectivity category)