Free
For getting started.
$0/ forever
Try the demoNo credit card required
- 5 deep assessments per month
- Unlimited summaries and key points
- All 10 CRAF dimensions
- Automatic source detection
- Limited Labs conversations
Powered by CRAF 4.6 Assess research.
Assess research.
See the
evidence.
Syllogos is an AI app for assessing research papers. See methods, strengths and weaknesses, with evidence you can check.
Three real papers. No signup.
SYLLOGOSCRAF 4.6
FULL ASSESSMENT
Beyond the breakthrough.
Understand the strengths.
Keep the limitations in view.
Methodological rigor
Originality & contribution
Objectivity & reflexivity
01
Explore a sample assessment Strong experimental design.
Limited reporting of uncertainty.
Built for critical reading10 dimensions Clear evidence Your judgment
Beyond a summaryRead the paper.
Read the paper.
Assess its quality.
Read, assess and organize your research in one app.
Start with your source.
Capture a screen, upload a file, enter a DOI or paste text. Syllogos retrieves available publication details and full text.
- Screen, file, DOI or pasted text
- Publication details and available full text
- You control capture and retrieval
What’s included
Capture a screen, upload a file, enter a DOI or paste text. Syllogos retrieves available publication details and full text.
- Screen, file, DOI or pasted text
- Publication details and available full text
- You control capture and retrieval
Understand the score.
Assess a paper with CRAF 4.6. Review its methods, originality, sources and bias, with evidence explaining the findings.
- 10 dimensions, 5 quality bands
- Evidence, strengths and limitations
- Focused or full assessments
What’s included
Assess a paper with CRAF 4.6. Review its methods, originality, sources and bias, with evidence explaining the findings.
- 10 dimensions, 5 quality bands
- Evidence, strengths and limitations
- Focused or full assessments
Syllogos
Assess as you read.
Open Syllogos over a paper, capture what you need and return to reading. Keyboard shortcuts keep it close.
- Desktop overlay and full-screen views
- Capture and analysis shortcuts
- Adjustable detection and appearance
What’s included
Open Syllogos over a paper, capture what you need and return to reading. Keyboard shortcuts keep it close.
- Desktop overlay and full-screen views
- Capture and analysis shortcuts
- Adjustable detection and appearance
Question the assessment.
Discuss a paper through five research perspectives, from Positivist to Pragmatist. Ask questions and challenge the reasoning.
- Five research paradigms
- Discussion grounded in the source
- Room to challenge the findings
What’s included
Discuss a paper through five research perspectives, from Positivist to Pragmatist. Ask questions and challenge the reasoning.
- Five research paradigms
- Discussion grounded in the source
- Room to challenge the findings
Keep your research together.
Save assessments alongside their sources. Revisit the evidence, compare papers or export a report.
- Saved reports and history
- PDF, Markdown and clipboard exports
- Source comparison
What’s included
Save assessments alongside their sources. Revisit the evidence, compare papers or export a report.
- Saved reports and history
- PDF, Markdown and clipboard exports
- Source comparison
See patterns in your research.
Explore disciplines, quality scores and recurring ideas across your saved assessments.
- Research fields and quality patterns
- Confidence levels and keyword trends
- Date and discipline filters
What’s included
Explore disciplines, quality scores and recurring ideas across your saved assessments.
- Research fields and quality patterns
- Confidence levels and keyword trends
- Date and discipline filters
AI assesses. You decide.Get Syllogos
The framework · CRAF 4.6
Ten dimensions.
Explore how CRAF assesses research.
Example assessment
Read the paper Attention Is All You Need · 2017
Dimension 01 of 10
Methodological rigor
8.7Strong
Were the methods rigorous?
Supporting evidence
Thorough ablation study isolates the impact of individual architectural components.
Key limitations
Lacks a dedicated limitations section discussing specific failure modes.
How to strengthen it
Add a substantive discussion of the model's limitations on extreme sequence lengths.
Dimension 02 of 10
Scientific validity
9.2Exemplary
Does the evidence support the conclusions?
Supporting evidence
Extensive empirical validation across multiple tasks with rigorous hyperparameter ablation.
Key limitations
Does not report variance or confidence intervals across multiple random seeds.
How to strengthen it
Include standard deviations or statistical significance tests for the BLEU improvements.
Dimension 03 of 10
Originality & contribution
9.8Exemplary
What does this work add to its field?
Supporting evidence
Introduces a fundamentally new architecture dispensing entirely with recurrence.
Key limitations
None; the contribution is unequivocally field-redefining.
How to strengthen it
Not applicable for a field-redefining, foundational contribution.
Dimension 04 of 10
Theoretical consistency
8.8Strong
Do the ideas and arguments hold together?
Supporting evidence
Mathematical formulation of attention perfectly aligns with goals of reducing sequential ops.
Key limitations
Connection between attention heads and linguistic structures is observational.
How to strengthen it
Provide formal theoretical proof of why specific heads capture syntactic dependencies.
Dimension 05 of 10
Literature mastery
8.3Strong
Does it engage with relevant research?
Supporting evidence
Accurately synthesizes limitations of existing recurrent models to motivate the new design.
Key limitations
The literature review is relatively brief regarding linguistic motivations for attention.
How to strengthen it
Expand background to critically evaluate prior work on hierarchical attention models.
Dimension 06 of 10
Source credibility
8.6Strong
Are the sources dependable?
Supporting evidence
Cites foundational and state-of-the-art papers from top-tier venues to establish baselines.
Key limitations
Heavy reliance on arXiv preprints bypasses formal peer review at the time of citation.
How to strengthen it
Update citations to their formally published versions where available.
Dimension 07 of 10
Writing quality
9.5Exemplary
Is the research clearly explained?
Supporting evidence
Exceptionally clear technical descriptions make a complex novel architecture easy to implement.
Key limitations
Some figure captions are brief and rely heavily on the main text for context.
How to strengthen it
Expand figure captions to be fully self-contained explanations of the visual diagrams.
Dimension 08 of 10
Peer evaluation status
8.7Strong
What peer scrutiny has it received?
Supporting evidence
Published in NeurIPS 2017, a universally recognized top-tier venue in machine learning.
Key limitations
Conference proceedings typically do not publish revision histories or reviewer notes.
How to strengthen it
Publish open peer review reports alongside the final manuscript.
Dimension 09 of 10
Objectivity & reflexivity
6.9Adequate
Does it acknowledge bias and limitations?
Supporting evidence
Maintains an objective register and transparently reports negative results in ablations.
Key limitations
Lacks explicit conflict-of-interest or funding disclosure statements.
How to strengthen it
Add explicit funding and conflict-of-interest declarations to meet modern standards.
Dimension 10 of 10
Implicit legitimacy
8.4Strong
What assumptions and authority does it rely on?
Supporting evidence
Provides an exceptionally detailed author contribution statement assigning specific credit.
Key limitations
Does not explicitly discuss the broader societal implications of scalable language models.
How to strengthen it
Include a brief discussion on the potential downstream impacts of the technology.
How to read the scoresPoor to Exemplary
Poor0–3.9
Weak4–5.9
Adequate6–7.4
Strong7.5–8.9
Exemplary9–10
Rubrics and calibration anchors set the scores. Checklists audit the evidence, rather than calculate scores. The overall score is weighted by discipline.
Inside the app
Your research. Six useful views.
Explore six views of the app.
01 / 06 · Add a source
Changes every 3 secondsStart with the paper you’re reading
Open Syllogos over a paper. Paste text, enter a DOI, upload a file or capture your screen.
FILM 01/Product overview
How Syllogos works.
See how the framework turns a paper into an assessment you can question.
Syllogos / Overview02:50
Research assessment in under three minutes.
4K product filmFilm 01Original overview · Includes earlier app screens
Sound on · Fullscreen available
Planned pricingStart free.
Start free.
Assess more.
Plans for individual researchers and teams. Paid plans are coming soon.
Recommended
Researcher
For regular research.
$18/ month
Coming soon$180/year · save two months
- 25 deep assessments per month
- Books and long-form sources
- Referee reports + more Labs
- Voice chat and read-aloud
- Saved research and full history
Pro
For larger projects.
$49/ month
Coming soon$499/year
- Higher deep-assessment allowance
- Batch papers and books
- Unlimited referee reports
- Priority deep-model access
- Everything in Researcher
Team / Lab
For teams and labs.
$22/ seat / month
Coming soonFrom $22 per seat · pooled usage
- Pooled deep assessments
- Shared feed and reports
- Workspace administration
- Centralized billing
- Institutional options
Flexible allowances. Extra deep assessments: $0.45 each.
Your judgment comes firstEvidence first.
Evidence first.
Your call.
AI assessments can make mistakes.
Check the evidence and apply your own expertise.
About SyllogosCommon questions
Download SyllogosGet Syllogos.
Get Syllogos.
For Mac & Windows.
Assess your own sources with CRAF 4.6 on Mac or Windows.
Try the browser demoChoose your platform
- 01Install
Windows: install from Microsoft Store. Mac: open the DMG and drag Syllogos into Applications.
- 02Add a source
Upload a paper, enter a DOI, capture a screen or paste text.
- 03Assess the paper
Run an assessment and review the evidence.
10 dimensions. Evidence behind the scores.Need help? Contact us ↗