Numbers with no source
Models produce precise-looking sample sizes, coefficients and p-values. When your supervisor asks where 0.594 came from, you are back to page one — or worse, there is no source at all.
WriterOne searches the elite repositories of your target country, extracts structured evidence cards, checks each number against the exact sentence it came from, and delivers the evidence appendix, comparison matrix and chapter draft your supervisor can verify.
Empirical study on factors influencing purchase decisions of Indonesian female fast-fashion consumers
Nilai R² = 0,594 menunjukkan bahwa faktor harga memiliki kontribusi paling besar terhadap keputusan pembelian konsumen perempuan di Indonesia.
None of these show up while you draft. They surface at the worst possible moment.
Models produce precise-looking sample sizes, coefficients and p-values. When your supervisor asks where 0.594 came from, you are back to page one — or worse, there is no source at all.
CNKI, Garuda, Neliti, SINTA, VJOL, ThaiJO, and the national statistics offices and central banks — general academic search indexes none of them, yet they are exactly the sources you most need to cite for your target market.
Your handbook wants a critique, coverage of the thesis focus, and at least one empirical study. A tool that returns a reading list gives you the wrong shape of work.
Not a summary of a summary. Every card in your library cleared deterministic checks before it was allowed in.
Only the target country's elite universities, indexed journals and authoritative agencies enter the pipeline. Anything else is dropped at search time, before it costs you an extraction.
Every number is stored together with the exact sentence it appeared in. If the sentence cannot be found, or the number does not match it, the evidence is discarded.
A statistic the source never reported is written as "not reported" — never filled in. With fewer than two verifiable data points the chart degrades into an explicitly labelled quality-audit chart, never a chart that merely looks like data.
Each step writes its own audit trail, so your supervisor can follow how every number arrived.
Fill 13 research-design fields, or upload your proposal and let AI extract them. From then on the system knows your phenomenon, your boundaries and your questions.
Routing is automatic: target country plus language gives you that country's elite sources, each with its tier and the reason it was included.
University theses, academic journals and authoritative reports run in parallel, followed by a second adaptive recall wave. Every query and every exclusion is recorded.
Cards are extracted, verified, tagged, scored for coverage and gap, and compared side by side across studies.
Cards are sampled by research dimension and argument role, then used to draft the Chapter 2 critique or the Chapter 3–4 triangulation — followed by a number-traceability re-check.
Everything below is generated from the verified card library and exportable to Word, Excel, Markdown and a self-contained audit page.
Method and sample, theoretical model and variables, reliability and validity, data-quality ledger, anchored sentences, GB/T 7714 and APA citations.
A 22-column ledger: metric definition, unit, period, geography, source, sample, comparability.
Researcher, sample, framework, IV→DV, significance, implication for your study — as a standard three-line table.
Argument step → required evidence type → minimum count → current gap. Your to-do list, derived from your thesis.
Identified, deduplicated, screened, full-text reviewed, included — with the reason each item was excluded.
Chapter 2 literature review with an explicit critique and research gap, or Chapter 3–4 internal/external data triangulation.
A standalone HTML page per card: highlighted original sentences, SHA-256 fingerprint, credibility score and every gate check.
Native-language and international literature are searched separately, so a study of Vietnam can cite both Vietnamese-language and English sources without either side being drowned out. The whitelist actually covers 142 target countries; 14 core markets are listed here.
| Country | Native track | English track | Reachable sources (university / journal / agency) |
|---|---|---|---|
| China | Top100 universities + Hong Kong and Macao; CNKI / Wanfang / VIP | Leading US and European universities + SSCI core journals | 227 / 16 / 20 |
| Indonesia | Top20 universities; Garuda / Neliti / SINTA | Leading US and European universities + Indonesian English-language journals | 138 / 14 / 20 |
| Vietnam | Top20 universities + the VJOL national database | Leading US and European universities + Vietnamese English-language literature | 128 / 14 / 13 |
| Thailand | Top20 universities + the ThaiJO national database | Leading US and European universities + Thai English-language literature | 127 / 14 / 13 |
| Malaysia | 24 verified universities + statistics office / central bank | Leading US and European universities + international journals | 142 / 12 / 14 |
| Philippines | 23 verified universities + central bank / statistics office | Leading US and European universities + international journals | 141 / 11 / 13 |
| Singapore | 30 verified universities + statistics office / monetary authority | Leading US and European universities + international journals | 148 / 12 / 13 |
| Japan | 28 verified universities + Bank of Japan / Ministry of Internal Affairs | Leading US and European universities + international journals | 146 / 11 / 13 |
| South Korea | 24 verified universities + global Tier-1 institutions | Leading US and European universities + international journals | 142 / 11 / 11 |
| United States | 12 verified universities + Federal Reserve / federal statistics | Leading domestic universities + international journals | 130 / 15 / 12 |
| United Kingdom | 17 verified universities + the national statistics office | Leading domestic universities + international journals | 135 / 11 / 12 |
| Germany | 24 verified universities + the national statistics office | Leading US and European universities + international journals | 142 / 11 / 12 |
| France | 23 verified universities + the central bank / the national library | Leading US and European universities + international journals | 141 / 14 / 12 |
| Australia | 30 verified universities + the central bank / statistics office | Leading domestic universities + international journals | 148 / 12 / 13 |
The whitelist is maintained entry by entry, by country and by language, and covers 142 target countries. A source that is not on it never appears, from search through to citation.
One thesis is one delivery. You pay when the work actually gets done, not to rent a seat.
Free
no card required
One thesis, full retrieval, 5 evidence cards.
$39 / ¥299
per thesis
One thesis, run end to end, everything exported.
$139 / ¥999
per year
Three theses, with room for the paper season.
Writing-service firms and cohorts: volume seats from $299 per seat per month.
No. It produces verifiable evidence cards, a data appendix, an evidence matrix and chapter drafts — you decide what to keep and how to argue it. When a number cannot be traced, it says so instead of smoothing it over.
No. A shared server quota works out of the box; you can also plug in your own key in your browser if you prefer.
Every number is stored with the exact sentence it came from and checked word for word. The archived original is searchable — Ctrl+F for it yourself — and every file carries a SHA-256 fingerprint.
The whitelist covers 142 target countries, and every country runs both a native-language and an English track. The site lists 14 core markets: China, Indonesia, Vietnam, Thailand, Malaysia, the Philippines, Singapore, Japan, South Korea, the United States, the United Kingdom, Germany, France and Australia. Deliverables can be produced in 64 languages, including Simplified Chinese, English, Bahasa Indonesia, Tiếng Việt, ภาษาไทย, 日本語 and 한국어. Cited passages are always kept in their original language and never translated.
It is parsed and archived inside your own thesis, never shared with other users. Delete the thesis or a single card and the archived files go with it.
The coverage panel tells you where the gaps are by channel and research dimension, how many cards you are short, and lets you top up only the gaps without clearing what you already have.
Free to start, no card required.
Open the workbenchYour uploaded files and retrieved sources stay in your own account.