Documentation
CScope
An open-source AI paper review and revision tool for researchers.
Install and run →
Prepare the runtime, configure a model, and create the first paper project.
Read the workflow →
See what each of the eight stages checks and how one stage informs the next.
Projects and data →
Review input formats, access tokens, retention, and how project data is used.
Configure the system →
Review model, MinerU, search, storage, and compilation settings.
What it does
After you upload a paper and select a target conference, CScope organizes the paper structure and related research, checks conference formatting, methodology, experimental results, and consistency across the manuscript, then produces a full review, section-level suggestions, and a revision checklist ordered by importance.
Key features
Conclusions tied to evidence
When assessing novelty or research positioning, CScope provides evidence that authors can check against their paper or the literature found during search.
Analysis from multiple perspectives
CScope examines the research background and novelty, methodology, whole-paper logic, experimental results, and writing separately before combining them into a full review.
Detailed revision suggestions
Scores provide only an overall judgment. CScope gives more detailed revision directions to help authors improve method choices, consistency across the paper, and experimental design.
A separate workspace for each paper
CScope creates a dedicated workspace and access token for every paper. Paper files, analysis results, and model interaction records remain in the corresponding project directory and are removed after the retention period.