Built for researchers

Systematic reviews,
no more spreadsheets.

Screen references double-blind, extract data from PDFs with the source behind every cell, read each paper with a critical eye and translate with your field's glossary — in one place.

free plan · no card active researchers
JAMA · 10.1001 · 2024

Continuous Glucose Monitoring in Type 2 Diabetes

Lin, Harper, Tanaka et al.
In 428 adults with poorly controlled type 2 diabetes, continuous monitoring lowered HbA1c by −0.72% at 12 months versus the control group.
SC Sarah's note — compare with Whitman 2023
Screening done 54 included · 312 excluded
used by researchers at
Tools 01 / 05

Four tools. One review, from screening to translation.

Screening decides what gets in; synthesis extracts the data; analysis reads each paper with a critical eye; the translator takes the text into another language without losing your field's terms.

N° 01

PRISMA Screening

Import exports from PubMed, Scopus, Web of Science, Embase or Zotero. Duplicates flagged on import, double-blind screening with your team and an exclusion reason logged on every decision.

CGM in T2Dincluded
Pediatric reviewpopulation
In-hospital glycemiareview
see screening →
N° 02

Tabular synthesis

AI suggests extraction fields from your papers and fills in the table. Click any cell to see the passage highlighted in the PDF.

Lin 2024 · n = 428p. 4
Tanaka 2022 · HbA1c −0.81%p. 7
Whitman 2023 · 12 monthsp. 3
columns configurablesource per cellExcel·CSV·PDF
see extraction →
N° 03

Critical analysis

Drop in a PDF and AI reads the whole paper: splits it into sections, highlights critical passages and flags inconsistencies that only show up when you compare one section with another.

Abstract × Results: n = 428 vs. 412mismatch
Methods: blinding not describedcheck
Conclusion backed by the dataok
sections splitcross-section checksper-passage rewrite
read a paper →
N° 04

Domain-aware translator in beta

Translate PDF, DOCX or text with your field's glossary in two passes — translation and polish — with canonical terms preserved.

hemoglobina glicada → glycated hemoglobinglossary
ensaio clínico randomizado → randomized controlled trialglossary
doi:10.1001/jama.2024.0427intact
10+ fieldsAI trainedDOIs intact
technical translation →
Manifesto 02 / 05
We believe serious research doesn't need more new tools — it needs an environment where they talk to each other.
EQ The Latvs team · Florianópolis
How it works 03 / 05

How it works, in four steps.

Researchers describe a path that repeats across disciplines. Latvs was designed around it.

N° 01

Screen

Import the exports from your search and screen double-blind, with duplicates flagged and the reason for every exclusion on record.

N° 02

Synthesize

AI suggests the fields and extracts the data from each included PDF, with the source passage highlighted in every cell.

N° 03

Analyze

A critical read of each paper: passages that deserve attention and inconsistencies between sections, flagged by AI.

N° 04

Translate

Take the text into another language with your field's glossary, without losing technical terms or citations.

14×
faster on initial screening vs. a manual spreadsheet.
8×
faster on data extraction vs. filling it in by hand.
−98%
of the cost per translation vs. a specialized professional.
Voices from our users 04 / 05

Researchers who stopped duplicating spreadsheets.

"
Screening that used to take me three weekends now fits in an afternoon. And every decision stays audited.
AC
Ana Carolina Sousa
PhD candidate in Public Health · UFMG
"
The domain-aware translator is what won me over. "MCG" comes out the way it should — and the project glossary beats the system one.
RT
Prof. Rafael Tanaka
Graduate program in Education · USP
Pricing 05 / 05

One plan, the whole review.

Start free, no card. All four tools on every plan — what changes is volume and support. Starting at R$ 39/mo.

Honest questions

Before you begin.

Does the AI make up data?

Every extracted value comes with the passage that backs it, highlighted in the PDF itself — you check it in one click before using it. The AI speeds up the reading; the call is still yours.

Where do my PDFs and data live?

Encrypted at rest and in transit, on audited providers. GDPR-aligned — and, the part that matters: we never train models on your material. References export as RIS/BibTeX, syntheses as Excel/CSV and translations as Markdown, and you can delete your account whenever you want. Details in the privacy policy.

Does it work for any research field?

Yes. Health, Social Sciences, Humanities, Law, Engineering, Education, Management, Psychology, Linguistics and Computer Science — among others. Review methods like PRISMA, ENTREQ and SPIDER and the per-discipline translator were designed to serve any academic field. Additional fields can be opened on institutional request.

Can I import my Zoterolibrary?

Yes. We import RIS, BibTeX and CSV files exported from Zotero, preserving the tags and notes on each reference. Export returns the same formats (RIS, BibTeX, CSV) — no data lock-in.

Does it replace my advisor?

Not at all. Latvs speeds up the mechanical work — screening, data extraction and technical translation — so you spend more time on what matters: interpretation and writing. The judgment stays yours.

Start your next review today.

Start free, no credit card. Import your references and run the review — screening, synthesis, analysis and translation — in one place.