Colandr

What Colandr is

Colandr is a free web platform for evidence reviews and other document reviews and syntheses. It helps review teams screen citation titles and abstracts by ranking records for relevance with machine learning. As reviewers include documents, natural language processing can suggest classifications for user-defined categories and support full-text data extraction. Reviewers retain the final say over which records to include or exclude and how to classify them, and team members can collaborate and cross-check decisions.

To begin, users bring citations from another system: Colandr does not connect to citation databases, though it can help generate Boolean search strings. Citation imports are supported as TXT, RIS, and BIB files, with a 40 MB maximum per import. Full-text papers must be located and uploaded individually. Screening decisions and extracted data can be exported as CSV. Relevance rankings refresh after every 10 citations are screened; metadata extraction suggestions start after 50 hand-coded extractions in a review. Colandr describes itself as open-source and open-access, with front-end and back-end code repositories. The community site says it runs on an AWS server hosted by Conservation International. The team directs users to a help email and bug-reporting form; volunteers run the project and ask users to allow time for replies.

Who it suits

Colandr suits researchers and review teams conducting evidence reviews or other document syntheses who want machine-learning support for citation screening and data extraction. It is also a fit for teams that need collaborative review and cross-checking while keeping final decisions with reviewers.

Limits

The free plan is listed as free. Citation imports are capped at 40 MB per import, full-text papers must be added one at a time, relevance rankings refresh after every 10 citations screened, and metadata extraction suggestions begin after 50 hand-coded extractions in a review.

Getting started

1. Open the Colandr website.
2. Export citations from another system, or use Colandr to help generate Boolean search strings.
3. Import citations as TXT, RIS, or BIB files, up to 40 MB per import.
4. Screen citations with relevance rankings and make inclusion or exclusion decisions.
5. Locate and import full-text papers individually for review.
6. Export screening decisions and extracted data as CSV.

Verdict

Choose Colandr for a free, web-based evidence review workflow with team collaboration and machine-learning assistance for screening and extraction. It may be less suitable if you need direct database connections or batch full-text uploads, since citations must come from another system and full texts are added one at a time.