GetlyЗаказная работа
Предложения заказной работы от Muse AI Lab
English to Japanese App UI Localization — Up to 500 Words
Launching your app or SaaS in Japan? Machine translation gives you Japanese words. We give you a localization file you can actually ship. We localize your English UI text into natural Japanese — and check the things raw translation always misses: button character limits, polite vs. casual tone, term consistency across screens, date and number formats, and placeholders that break when the word order changes. Starter package — what's included - Up to 500 English words / 100 strings, one app, one feature area - English → natural Japanese localization (not word-for-word translation) - Tone unified to your product's voice (polite desu/masu by default) - Term consistency checked across all strings - Character limits respected where you provide them - Placeholders and variables ({name}, %s, ICU plurals) preserved and repositioned correctly - Delivered in ONE format of your choice: UTF-8 CSV, JSON, or XLIFF 1.2 — same keys, same order as your source - QA report included: checks performed, decisions made, and items that need your confirmation Not included - Code implementation or in-app testing on real devices - Text inside images, legal translation, or app-store review guarantees - Marketing copy / landing pages (ask first — scope differs) What we need from you when you order 1. Your source file (CSV / JSON / XLIFF with the English strings) 2. What the app does and who it is for 3. Your target users in Japan (business / consumer / gamers…) 4. Preferred tone (polite / friendly / brand-specific), if any 5. Character limits per string, if your UI has them 6. Delivery format: CSV, JSON, or XLIFF 1.2 About confidentiality If your strings are from an unreleased product, tell us before ordering. We will explain exactly how your text is handled and get your OK before any of it is processed with external tools. Sample: the gallery shows a fictional 30-string sample (a task app) with its QA notes — the same format you receive.
CSV Rescue — Clean & Fix Broken Import Files
Your CSV import failed again. Encoding errors, shifted columns, dates your system refuses, duplicates you can't trust. Send us the broken file. We clean, validate and reformat it to your target system's spec — deterministically, with tools and rules, not guesswork. No value is ever invented or "filled in" by an AI. What we fix - Character encoding (including mojibake / garbled text) → clean UTF-8 - Delimiters, quoting and line-ending problems - Date and number formats unified to your target format - Column headers renamed and reordered to your spec - Exact and near-duplicate rows removed (reported, never silently dropped) - Validation against rules you supply (required fields, formats, ranges) Scope - 1 CSV file, up to 1,000 rows and 20 columns - Deliverables: cleaned CSV + a change log listing every fix + a short validation report (rows fixed, rows flagged, rows we could not repair and why) - Your original file is never overwritten. Unrepairable rows are reported, not hidden. What we don't do - We never invent missing data or guess values - No complex database migrations, no access to your live systems - We cannot accept files containing personal data, medical or payment information, passwords or private keys. Please anonymize sensitive columns before ordering. What we need from you when you order 1. The CSV file to fix 2. The target system you will import into (e.g. Shopify, WooCommerce, QuickBooks, your CRM) 3. The column layout / format you need (a template or sample row is ideal) 4. Any error messages your system showed, if you have them Data handling Your file is used only to complete your order. Our working copies are deleted within 30 days after delivery. Records kept by Getly follow Getly's own policy. Questions before ordering? Send a message — we reply with whether your file is in scope before you pay.
Website & Landing Page Audit — GPT + Claude
Find overlooked issues before your users do. Launching a landing page, website, or small web app? Before you go live, get a structured review of your product from two independent AI systems: OpenAI's GPT and Anthropic's Claude. Each AI reviews your materials independently before their findings are compared and challenged. The final report highlights potential issues, explains why they matter, and gives you a prioritized list of practical improvements. You don't just get feedback. You get a clearer picture of what to fix first. WHO THIS IS FOR - Indie hackers preparing to launch a product - Solo founders building landing pages or small SaaS products - Freelancers reviewing a website before client delivery - Small teams looking for a second perspective on usability, messaging, and user experience WHAT'S INCLUDED One landing page OR up to five screens of a small web product, plus up to 5,000 Japanese characters (or approximately 2,500 English words) of supporting documentation. We review clarity & messaging, user experience, conversion flow (CTAs, pricing explanations, signup expectations), consistency (contradictions, ambiguous claims, missing details), user trust (unsupported claims, unclear data handling), and where the two independent reviews agree or disagree. Every finding includes what the issue is, why it matters, its potential impact, a practical recommendation, and an evidence classification (Confirmed Textual Issue / Reasonable Concern / Hypothesis / Not Verified). WHAT YOU'LL RECEIVE A structured 7–10 page PDF report: executive summary, review scope & criteria, GPT independent findings, Claude independent findings, agreements/disagreements & cross-validation, prioritized issues, recommended improvements, and limitations & unverified areas. Delivery: within 3 days (72 hours) of order placement, provided all required materials are submitted in the order brief and the review is within the stated scope. If materials are missing or the scope needs clarification, we will contact you promptly and agree on a revised delivery date where necessary, subject to Getly's order and delivery rules. A NOTE ON OUR SAMPLE TEST In our internal sample test, both AI systems together identified all 10 predefined issues. One model independently identified 9, while the other identified all 10. The cross-review also helped refine findings and their evidence classifications. This was a single controlled demonstration, not a statistical benchmark. Agreement between two AI systems does not guarantee that a finding is correct. WHAT'S NOT INCLUDED This is an AI-assisted pre-launch review, not a certification or professional audit. It does not include security testing or guarantees, legal or regulatory compliance certification, guaranteed conversion or revenue improvements, source-code auditing, live usability testing with real users, or implementation of recommended changes. PRIVACY & DATA HANDLING Your materials are used only to perform and deliver your review, and are processed by OpenAI and Anthropic AI services for the independent audits and cross-check. We delete our working copies of your submitted materials within 30 days after delivery. Order records stored by Getly are handled under Getly's privacy policy; we cannot delete platform-side records ourselves. We never publish your materials or use them in examples without your permission. Please do not submit passwords, sensitive personal information, confidential customer data, or materials you are not authorized to share. FAQ Q: How is this different from asking ChatGPT to review my website? A: Two different AI systems independently examine your materials before their findings are compared and challenged. You receive a structured, prioritized report rather than an unorganized chat response. This method can uncover different perspectives, but it does not guarantee more accurate results. Q: Do I need a live website? A: No. Screenshots, wireframes, and PDFs are accepted. Please make sure the materials clearly show the screens you want reviewed. Q: Can you review my entire app or codebase? A: Not under this package. This review covers one landing page or up to five selected screens, with limited supporting documentation. We do not review source code or test backend functionality. Q: What if the review identifies no major issues? A: You will still receive an assessment of the submitted materials, including their strengths, limitations, and recommended next steps where applicable. We do not invent issues to fill the report. Japanese-language websites and materials are welcome. 日本語のLP・Webサービスも監査対象です。日本語レポートでの納品にも対応します。