This work exploits the use of a fuzzy many-valued concept analysis for document retrieval to efficiently answer users' queries. For the purpose of visualization, we propose to deal with fuzzy many-valued formal contexts in a tripartite graph with respect to fuzzy triadic document co-similarities. The proposed model aims at describing the documents according to three hierarchical levels. It is based on a prototyping of feature vectors applied to documents, sentences and words. The recourse to fuzzy conceptual scaling has enabled us to visualize many-valued concepts through the construction of fuzzy nested lattices. Once this built, we proceed to document retrieval using building process and lattice incremental path such a simple information retrieval.