Find Source Code for Research Papers
Automatically find GitHub repositories and source code implementations for academic papers. PapersFlow extracts code references, searches GitHub, and lets you inspect repos without leaving your research workflow.
Paste a paper DOI or URL, and PapersFlow finds the associated GitHub repositories, source code, and datasets -- helping you reproduce results and build on existing implementations.
You read a paper with promising results and want to reproduce them or build on the method. So you start hunting for the code. You check the paper for a GitHub link — sometimes it's there, sometimes it's broken, sometimes it doesn't exist. You search GitHub manually but the repo name doesn't match the paper title. You find three repos and aren't sure which is the official one. This process takes 15-60 minutes per paper and often ends in frustration. Reproducibility shouldn't require detective work.
Key Features
- Multi-Strategy GitHub Search
- In-App Repository Inspection
- Paper-Code Linking
- Availability Assessment
Tools
Compare
Frequently Asked Questions
- What if the paper doesn't have any published code?
- PapersFlow will tell you clearly when no code can be found. This is useful information in itself — it helps you assess reproducibility during literature reviews and know when you'll need to implement from scratch.
- Does it find datasets too?
- If a paper references datasets with URLs or well-known names (e.g., 'ImageNet', 'GLUE benchmark'), PapersFlow will extract those references. However, dataset discovery is secondary to code discovery — for comprehensive dataset search, specialized tools exist.
- Can I search for code during a literature review?
- Yes. Code Discovery is integrated with the chat interface. During any research session, ask 'Find the code for this paper' and PapersFlow will search without interrupting your workflow.
- How does it handle papers with broken GitHub links?
- PapersFlow doesn't rely solely on URLs in the paper. If a link is broken (common after repo renames or organization changes), it falls back to searching by paper metadata, author profiles, and code patterns to find the current location of the repository.