Rat models are frequently used for finding genes contributing to the arthritis phenotype. In most studies, however, limitations in the number of animals result in a low resolution. As a result, the linkage between the autoimmune experimental arthritis phenotype and the genomic region, that is, the quantitative trait locus, can cover several hundred genes. The purpose of this work was to facilitate the search for candidate genes in such regions by introducing a web tool called Candidate Gene Capture (CGC) that takes advantage of free text data on gene function. The CGC tool was developed by combining genomic regions in the rat, associated with the autoimmune experimental arthritis phenotype, with rat/human gene homology data, and with descriptions of phenotypic gene effects and selected keywords. Each keyword was assigned a value, which was used for ranking genes based on their description of phenotypic gene effects. The application was implemented as a web-based tool and made public at http://ratmap.org/cgc. The CGC application ranks gene candidates for 37 rat genomic regions associated with autoimmune experimental arthritis phenotypes. To evaluate the CGC tool, the gene ranking in four regions was compared with an independent manual evaluation. In these sample tests, there was a full agreement between the manual ranking and the CGC ranking for the four highest-ranked genes in each test, except for one single gene. This indicates that the CGC tool creates a ranking very similar to that made by human inspection. The exceptional gene, which was ranked as a gene candidate by the CGC tool but not in the manual evaluation, was found to be closely associated with rheumatoid arthritis in additional literature studies. Genes ranked by the CGC tools as less likely gene candidates, as well as genes ranked low, were generally rated in a similar manner to those done manually. Thus, to find genes contributing to experimentally induced arthritis, we consider the CGC application to be a helpful tool in facilitating the evaluation of large amounts of textual information.