TransLiG is an efficient de novo trascriptome assembler for RNA-Seq data. It can assemble all transcripts from short reads (single or paired) without using a reference. The software expects as input RNA-Seq reads in fasta or fastq format, and ouput all assembled candidate transcripts in fasta format. Briefly, it works in two step: first, TransLiG partitions the sequence data into many individual splicing graphs, each capturing the full transcriptional complexity at a given gene or no more than a few genes. Then, TransLiG repetitively phases and contracts paths over the splicing graphs, infering an optimal edge-path-cover for each graph representing the expressed transcripts of this gene locus. This software is free to use, modify, redistribute without any restrictions, except including the license provided with the distribution.