Users need more sophisticatedtools to handle the growing numberof image-based documents availablein databases. In this paper, wepresent a system devoted to theediting and browsing of complexliterary hypermedia includingoriginal manuscript documents andother handwritten sources. Editingcapabilities allow the user totranscribe manuscript images in aninteractive way and to encode theresulting textual representationby means of a logical markuplanguage (based on the XML/TEIspecification). Bothrepresentations (image andstructured text) are tightlylinked to facilitate the readingand the interpretation ofdocuments. This text/imagecoupling scheme is an attempt tounify several layers ofinformation in order to providethe user with a global vision ofthe work. Our system also suppliestools capable of processing andrelating information stored bothin images and structured texts.Finally, application-specificvisualization techniques have beendeveloped in order to provideusers with a way to identifyrelationships between sourcedocuments and help them tonavigate.