Two weeks of literature work, in two hours

Multi-database search, journal tiering, appraisal and a first draft in one pass — every citation traceable.

Create account — 3 free uses See a real output
PubMed · Scopus · OpenAlex · 7,499 journal templates Hosted in China · manuscripts never shared · AI content labelling

AI first draft — human verification required · generated content is labelled as required by China's AI content labelling rules

Real output example
  • 120 of 221records included
  • 19sections
  • 67/67citations verified

Downloads: Word / Markdown

Six entry points, one pipeline

Credits are consumed per task and always visible in your account panel. Modules work independently.

Core

Literature review

Multi-database search → journal tiering → appraisal → quick/deep draft → citation verification → export.

CreditsSearch 1 credit; quick draft +1; deep draft +3 (outline step free). Not charged when citation verification falls below 70%.

Open workbench
Inside the review module

Academic translation & polishing

Chinese ⇄ English academic conversion and same-language polishing; long texts are chunked with citation markers preserved.

Credits+1 per successfully processed chunk; failed chunks are not charged

Open workbench (translation)
Standalone

Three-line tables

Build journal-formatted tables from your review content and data blocks; shared with the slide module.

Creditsrendering costs 2 credits (producing the table plan is free)

Open three-line tables
Standalone

Presentation slides

Turn the review and tables into a slide deck (12 page types × 8 visual themes) with speaker notes.

Credits2 credits

Open slides
Compliance · Free

AI-labelling check

Scans Word/text files for AI-related metadata, unusual invisible characters and mixed-script homoglyphs, and returns item-by-item fix guidance.

Freesign-in required; detection only — we do not offer label removal

Open AI-labelling check
Free

Analysis workshop

Clinical statistics and publication figures: 50+ calculators, 30+ scientific plot types, Western blot quantification — computed locally and reproducible.

Freesign-in required; exports carry a source mark

Open analysis workshop

Paper packs & credits

Priced by paper, one-off purchase, no auto-renewal; credits can be split across search, statistics and figures. 3 free credits on registration (valid 30 days); the two free tools consume no credits.

Pack · Small
10 papers

~20 credits

Trial a topic or one revision round

Pack · Medium
30 papers

~60 credits

One full review (search · draft · verification · statistics · revision)

Pack · Large
100 papers

~200 credits

Continuous use for a research group

How are credits counted? (open for the conversion)
Priced by paper, metered in credits: 30 papers ≈ 60 credits (about 2 credits per paper). Deep mode costs about 4 credits per paper; statistics and figures are metered separately, and each task shows its cost before it runs. A generation that fails the 70% citation threshold is not charged. The old per-credit plans are discontinued; existing credits were migrated.

Price is confirmed one-to-one based on length, field and deadline — just fill in the form below and you will get a quote within 24 hours (mention invoices or team/department sharing in the message).

A reproducible pipeline: the machine does the repetitive work, the judgement stays with you

①

Search

Multi-database search with automatic de-duplication.

You confirm: is the query on target

②

Tier

SJR quartile mapping to T1/T2/T3, threshold adjustable.

You confirm: inclusion threshold

③

Appraise

7 study-type guides plus bias and reporting dimensions (RoB2, ROBINS-I, QUADAS-2, NOS, AMSTAR-2, GRADE, PRISMA).

You confirm: appraisal conclusions

④

Draft

Quick or deep mode; deep mode presents an outline for approval first.

You confirm: framework and inclusion

⑤

Verify

Citation existence and semantic consistency checks, with a traceability table.

You confirm: what to do with failed citations

⑥

Export

Word / Markdown / RIS / BibTeX / EndNote XML.

You confirm: final quality

Read the full guide

Three steps to start

1

Register — 3 free credits

Sign up with an email; credits are per account and valid 30 days.

No payment and no manuscript upload needed to start.

2

Enter your question, screen the hits

Chinese or English (Chinese is converted into a search query): multi-database search, SJR-quartile tiering, then tick the records you want to include.

The threshold is adjustable; tiering filters, you decide inclusion.

3

Generate, verify, export

Quick or deep mode produces a draft → review the citation check → export Word / Markdown / RIS / BibTeX / EndNote package.

Failed citations are flagged; whether to keep them is your call.

Create account — 3 free credits

AI can write fast. A medical review has to be checkable.

Per-citation verification

Every citation is checked against PMID/DOI records and for semantic consistency. Citations that fail are flagged at the end for your decision instead of being quietly left in the text.

The 70% threshold

If fewer than 70% of citations pass verification, that generation is not charged. We would rather redo it than let it through.

Evidence anchors

Each in-text citation attempts to locate the supporting sentence in the source (abstract or full text) and states the source level. When nothing can be located it is marked "not located" for your check — never filled in by the program.

Programmatic reference rendering

References are rendered deterministically from 7,499 EndNote journal templates — the model does not improvise. Export to RIS / BibTeX / EndNote XML.

AI content labelling

In the interface and in exported files, AI-generated content is labelled and written into document metadata. Please do not remove or obscure it before redistribution.

What we do not do

No ghostwriting, no rewriting to evade plagiarism checks, no fabricating or cosmeticising data, no promise of acceptance; and no label removal.

Output is an assisted first draft. It does not replace the formal conduct of a systematic review or meta-analysis, and it is not medical advice.

What happens to your unpublished data

No third parties

Manuscripts and data are not fed into any third-party training or evaluation pipeline.

30-day retention

Files and outputs from self-service tasks are deleted automatically after 30 days; manuscripts handled in human services are deleted within 30 days of delivery. Ask us any time for earlier deletion.

Hosted in China

Served from a mainland China cloud host, with ICP filing (湘ICP备2026006398号-2) and public-security filing.

See what it actually produces

Examples are taken from real outputs and anonymised; verification sheets and exports are downloadable.

① Article body (exported Word)
① Article body (exported Word)Page 1 of the exported Word file — title and abstract laid out in the target journal's format.
② Citation verification report
② Citation verification reportThe same sample after verification: 67 citations checked, 67 passed, 0 pending manual review. Each row states its source level.
③ Reference rendering
③ Reference renderingRendered programmatically from journal templates (not improvised by the model), with PMID/DOI kept for manual checking.

以上三张均取自公开示例文件（example-en.docx / example-en.md），未修改内容，仅截取局部；原件可在 /samples 下载。

Case A

Clinical question → deep-mode review

120 of 221 included · 19 sections · 67/67 citations verified

Includes excerpt + citation verification sheet

Case B

Research topic → quick-mode summary

~500-word condensed summary with citation markers and reference list

Includes export

Case C

EndNote submission package

Target-journal reference style plus RIS / BibTeX / EndNote XML

Includes sample files

Browse all samples and verification sheets

Frequently asked

What happens to my unpublished manuscript?
Manuscripts and data are never uploaded to third-party platforms and are not used for model training. Self-service outputs and uploads are deleted automatically after 30 days; human-service manuscripts are deleted within 30 days of delivery. The service runs on a mainland-China host with ICP and public-security filings.
Is this ghostwriting?
No. We run the search, tiering, appraisal notes, draft and formatting; the topic and authorship stay yours. No ghostwriting, no plagiarism-evasion rewriting, no fabricated or cosmeticised data, no acceptance promises.
Can the citations be made up?
Citations come from PubMed / Scopus / OpenAlex. After generation every citation is checked for PMID/DOI existence and semantic consistency; failures are flagged for your decision, and references are rendered programmatically from 7,499 journal templates.
How are credits counted?
Priced by paper, metered in credits: search 1, quick draft +1, deep draft +3 (outline free), three-line tables 2, slides 2, translation +1 per successfully processed chunk. A generation failing the 70% citation threshold is not charged. Registration includes 3 free credits valid 30 days.
Can I submit to an English-language journal?
Yes. Bilingual generation and same-language polishing are supported, with export to Word / Markdown / RIS / BibTeX / EndNote XML and references rendered in your target journal's style.
Invoices, or sharing credits across a team?
The service is currently handled directly by the author; use either form on this page to describe your need and we will propose an arrangement.

Self-service tools and human services

Self-service (uses credits)

  • Literature reviewTranslation & polishing
  • Three-line tablesPresentation slides
  • AI-labelling check (free)Analysis workshop (free)

Priced by paper, metered in credits; see your account panel for details

Human services (by our team)

  • Manual language polishing · Expert consultation
  • Study design · Troubleshooting

No ghostwriting, no plagiarism-evasion rewriting, no data beautification, no acceptance promises.

See human services

Register, and hand the next literature search to the pipeline

Create account — 3 free uses

Team or department use, invoices, or custom delivery: eyuanshangou@163.com