RU Join the waitlist

Research made simpler, faster, open to all

Sources, data, text and checks – in one place. Scipen takes a clinical study from the spreadsheet to the journal article: AI agents match every citation against its source, every number against your data, and take the routine off your hands.

Sources

Search PubMed, Europe PMC, CrossRef and Russian databases right from the project. Every paper shows whether the full text is available, and CitationHub keeps verbatim quotes and finds them by meaning – in any language.

Sources
early laparoscopic cholecystectomy, patients over 70
  • Early laparoscopic cholecystectomy in patients aged 80 and over: a multicentre cohort2022 · PubMed · cited in the text as [19]
    full text
  • Laparoscopic versus open cholecystectomy in the elderly: a systematic review and meta-analysis2021 · PubMed
    full text
  • Frailty and outcomes after emergency cholecystectomy2023 · PubMed
    abstract only
In the text 31On topic 12Set aside 3example

CitationHub · verbatim quote [19]

by meaning: “lower 30-day mortality”

…30-day mortality was lower after early laparoscopic surgery (RR 0.62; 95% CI 0.41–0.94), while major morbidity did not differ.

Data

Your Excel sheet opens inside the project: variables, missing values, analysis. Table 1, forest plots and Kaplan–Meier curves come out journal-ready, and missing data are imputed with MICE.

Data · sheet “Cohort”example
Table 1. Patient characteristics
CharacteristicLC, n = 268OC, n = 144p
Age, years, median [IQR]75 [71–80]78 [73–83]0.002
Women, n (%)171 (63.8)88 (61.1)0.59
ASA III–IV, n (%)97 (36.2)71 (49.3)0.01
Tokyo II–III, n (%)112 (41.8)79 (54.9)0.01
BMI17 of 412 · 4.1%
Albumin40 of 412 · 9.7%
ASA2 of 412 · 0.5%

MICE: 20 imputed datasets, estimates pooled with Rubin’s rules. The original sheet stays untouched – we work on a copy.

Text

Numbers in the manuscript are linked to cells in your data. When the sample changes, Scipen flags the stale numbers and updates them with one click.

Results linked to data

The study included 412 patients: 268 (65.0%) underwent laparoscopic and 144 (35.0%) open cholecystectomy. Clavien–Dindo grade III or higher complications occurred in 6.3% of the LC group and 14.6% of the OC group.

The data changed: 2 patients were excluded from the OC group (readmission). Four numbers in the text are out of date.

Checks

Every citation is matched against the text of its source, every number against your data. The source itself is checked for existence: DOI and PMID resolved again, retracted papers flagged.

In the textEvidenceVerdict

[14] “laparoscopy in patients over 75 means fewer severe complications”

Source: “lower rate of major complications in patients over 75”

confirmed

[19] “a threefold reduction in mortality”

Source: “RR 0.62; 95% CI 0.41–0.94” – a reduction of about 38%

contradicts

“8 days” – median length of stay, OC

Table 2 and the data: 9 days

mismatch

[27] doi:10.1007/s00464-0…

CrossRef: DOI not found; PubMed: no record

not found

One citation, three states

This is how Scipen shows every citation: what came in with the manuscript, what the source actually says, and what it became after your decision.

As submitted

Several authors report a threefold reduction in mortality with early laparoscopic surgery [19].

not checked

What Scipen found

Several authors report a threefold reduction in mortality with early laparoscopic surgery [19].

Source [19] found · PMID verified · full text

…30-day mortality was lower after early laparoscopic surgery (RR 0.62; 95% CI 0.41–0.94), while major morbidity did not differ.

Contradicts: the source reports a reduction of about 38%, not threefold.

Several authors report a threefold reduction in mortality a reduction in 30-day mortality (RR 0.62; 95% CI 0.41–0.94) with early laparoscopic surgery [19]

contradicts

After the edit

Several authors report a reduction in 30-day mortality (RR 0.62; 95% CI 0.41–0.94) with early laparoscopic surgery [19].

confirmed

Example. The manuscript, source and numbers are illustrative.

What it looks like inside

A project is open on the desktop. The manuscript is dragged into the window, Scipen checks the citations and opens the source, Supervisor suggests an edit, and the author applies it.

Example: the manuscript and source are illustrative.

AI suggests. You decide.

Supervisor sees the whole project: manuscript, data and sources. It searches the web, suggests edits and sets tasks. You apply or reject them, and the run log keeps a record of what every agent did.

Supervisor · suggestions3 awaiting decision

Abstract

The abstract is over the journal limit: 274 words against 250.

Remove the repeated methods and cut the conclusion to one sentence: 248 words.

AppliedRejected

Methods

Handling of missing BMI values (4.1% of observations) is not described.

Missing BMI values were imputed by multiple imputation (MICE, 20 datasets).

AppliedRejected

Results

The median length of stay in the OC group does not match Table 2.

The median postoperative length of stay was 5 and 8 9 days

AppliedRejected
Run log
  1. Citation check · 46 sourcesdone
  2. Number audit · 128 values, 2 mismatchesdone
  3. Web search · 3 new papers on topicdone
  4. Translation RU → EN · “Discussion”running

The model never learns who your patient is

Before anything goes to a cloud model, Scipen replaces names, medical record numbers, insurance and ID numbers, phone numbers and dates of birth with tags. The answer comes back to you with the real values, and a local model runs on your own computer – the data never leave it.*

On your side · sheet “Patients”as is
MRNNameDate of birthDOBGroupCD ≥ III
12841/24Anna S. Petrova14.03.1949LCno
13077/24Oleg I. Smirnov02.11.1946OCIII
13102/24Maria P. Orlova27.06.1952LCno
13240/24Igor A. Volkov09.01.1944OCII

Chat questionPatient Anna S. Petrova, tel. +7 916 555-12-34, was readmitted – should she be excluded from the analysis?

Sent to the cloud modelreplaced: 0
MRNNameDate of birthDOBGroupCD ≥ III

Chat questionPatient , tel. , was readmitted – should she be excluded from the analysis?

*Built with personal-data law (Russian Federal Law 152-FZ) and medical confidentiality in mind: patient details never reach cloud models in plain form.

Journal, rules, letter, file

A catalogue of 581 journals: SJR Q1–Q3 plus the Russian VAK list. Scipen checks the manuscript against the chosen journal’s rules, drafts the cover letter and the response to reviewers, translates RU ↔ EN and exports the file.

Journal finder · 581

    Requirements: Surgical Endoscopy

      Requirements in this example are illustrative

      Translation RU ↔ EN5 of 5 intact
      RU

      Осложнения III степени и выше развились у 6,3% и 14,6% пациентов (ОШ 0,39; 95% ДИ 0,20–0,76) [14].

      EN

      Clavien–Dindo grade III or higher complications occurred in 6.3% and 14.6% of patients (OR 0.39; 95% CI 0.20–0.76) [14].

      Hover over a number: Scipen checks that every value and every citation made it into the translation unchanged.

      Exportready to submit

      Manuscript, tables and figures in DOCX in the journal’s style, references in its format.

      A journal that fits the work, not a guess

      Scipen looks at where your sources and similar papers were published, weighs the study design, sample size and registration, and sorts journals into three tiers. Each comes with time to acceptance, publication fees and what you’ll need to do. Preferences are optional: they only lift the journals that match.

      Journal match · LC in patients aged ≥70catalogue: 581 journals

      Your study

      • Retrospective cohort
      • n = 412
      • 2 centres
      • STROBE 22/22

      Where the signal comes from

      • Journals of your sources14
      • Similar papers, OpenAlex38
      • Catalogue: SJR and VAK581

      Preferences optional

        Example: the journals are real; scores, timelines and fees are illustrative.

        Your own key – or no key at all

        Each role gets its own model. Work with your own API key within a budget you set yourself, or plug in a local model – then you need neither a key nor a model subscription.

        Settings · models by roleown key: 2 · local: 2

        Supervisorchat, edits, tasks

        Citation checkmatching against the source

        Statisticsanalysis and results text

        TranslationRU ↔ EN with number checks

        Monthly budget41% of the limit used · you set the limit

        example

        Join the Scipen waitlist

        Send my invitation to 

        You are

        Scipen is in closed development. Access opens through the waitlist – those who sign up get it first.