Cross-platform Mobile QA Tester — Flutter iOS/Android/macOS/Web + RevenueCat

Keppt — Pre-Launch QA Engagement Engagement type: Fixed-price, milestone-based Timeline: 3–4 weeks Budget range: $1,500–4,000 fixed (higher tier for candidates with cross-platform mobile IAP experience) Client: Fixonal PTE. LTD. (Singapore) — indie mobile/SaaS studio THE PRODUCT Keppt is a cross-platform AI-powered document catalogue for personal paperwork. Users upload IDs, contracts, policies, medical records, warranties, tax forms, receipts. AI extracts fields (vendor, dates, amounts, reference numbers), files documents into Libraries/Collections, surfaces upcoming renewals and expiries, and answers plain-language questions with source citations. Live at: keppt.me (web app usable) Platforms: iPhone, iPad, Android, macOS, Web. Flutter monorepo, Firebase backend, Vertex AI Gemini for extraction, RevenueCat for subscriptions, Sentry for error tracking. Business model: Freemium subscription — Free (10 docs), Basic $4.99/mo, Unlimited $12.99/mo, Family $19.99/mo. Priced globally with regional variants. Constraint: Fixonal's brand is deeply privacy-first. Any bug that appears to leak or mishandle user data is treated as maximum severity. WHAT WE'RE HIRING FOR An experienced QA tester (or small team) to run a structured exploratory QA pass across all supported platforms before launch. Not automated test authorship — manual, deep, adversarial testing by a human trying to break real user flows on real devices. Critical hire criteria: - Demonstrable cross-platform mobile QA experience (iOS + Android at minimum, macOS and Web a strong plus). - Direct experience testing in-app purchase / subscription flows — RevenueCat, StoreKit, Google Play Billing, and any web billing stack. Subscription bugs are our highest revenue risk. - Comfort with AI/LLM output validation — verifying extracted document fields against source content and flagging hallucinations or off-by-one errors. - Ability to write actionable bug reports — repro steps, expected vs actual, device/OS/build/env, screenshots or short videos, severity + priority tagging. - Willingness to test with real personal documents (or produce a realistic synthetic test corpus) — the AI extraction can't be tested with lorem-ipsum PDFs. Not hiring for: automation test framework setup, security penetration testing (separate engagement), performance load testing, or UX critique. SCOPE — WHAT TO TEST (in priority order) 1. Subscription and paywall flows (highest revenue risk) — Purchase on iOS (App Store), Android (Play Store), and web billing; restore purchases across devices; upgrade/downgrade between tiers; cancel + expiry behaviour; refund handling (RevenueCat webhook path); Family plan sharing (up to 4 users); Free tier limit enforcement; Sandbox vs production Apple ID behaviour. 2. Cross-platform sync — Document uploaded on iPhone appears on Mac, iPad, Android, web; sync conflicts; offline behaviour on mobile; startup performance with large libraries. 3. AI extraction accuracy — Test corpus of 20+ real documents across categories; verify extracted fields against source; non-English documents; handwritten annotations; poor-quality scans, phone photos, dark PDFs; multi-page documents; documents with sensitive PII. 4. Auth and MFA — Sign in with Apple flow on each platform; TOTP MFA setup and recovery; password recovery, lost 2FA recovery; sign-out behaviour and re-auth. 5. Data residency and region behaviour — Verify data actually pins to user-selected region (US/EU/APAC); CN gate: signup from China IPs is blocked; GDPR export flow; GDPR delete flow. 6. Ask feature (LLM with citations) — Verify citations link to source documents; Ask doesn't hallucinate documents; behaviour when asked about docs not in library. 7. Platform-specific — iOS: VoiceOver, Dynamic Type, iPad multitasking, macOS Handoff; Android: back-button, deep linking, notification permissions; Web: browser matrix; macOS: menu bar, keyboard shortcuts. 8. Localization / i18n — Spot-check strings across supported languages; currency/date-format regionalisation. 9. Edge cases — Very large files, corrupted files, unsupported file types; rapid uploads; full library. DELIVERABLES (MILESTONE STRUCTURE) Milestone 1 — Test plan + bug tracker setup (Week 1, ~$400–800) Written test plan mapping every category above to concrete test cases with device/OS/build matrix; recommended bug tracker; severity/priority scheme; sign-off on scope from us before execution begins. Milestone 2 — Execution pass 1 (Week 2–3, ~$700–2000) Systematic execution of the test plan; bug reports filed in the tracker; weekly progress update; immediate flagging (same-day) of any P0. Milestone 3 — Regression pass + sign-off (Week 4, ~$400–1200) Retest of all P0/P1 bugs after we fix them; written pre-launch QA sign-off report including one-paragraph "would I launch this?" honest opinion. WHAT WE NEED FROM YOU IN YOUR PROPOSAL 1. QA background — years of experience, mobile vs web split, largest cross-platform product you've QA'd. 2. IAP / subscription experience — specific products where you found subscription bugs. RevenueCat / StoreKit / Play Billing familiarity. 3. Sample bug report — attach an example of an actual bug report you've written (redact confidential product names if needed). 4. Devices available — list physical devices you can test on. If limited, tell us how you handle the coverage gap. 5. AI/LLM validation experience — have you tested products with AI-generated content? 6. One question you have about Keppt after reading this brief — filter for candidates who actually engage. 7. Rate and milestone breakdown. HIRING APPROACH Fixed-price total, milestone-based, funded through Upwork escrow. Milestone 1 sign-off is a checkpoint. Async communication via shared bug tracker; one 30-min sync per milestone. All bug reports must land in the tracker. Proposals missing sample bug report, device list, or the Keppt question are auto-declined.