The role of endoscopic haemostasis in the treatment of bleeding peptic ulcers is widely studied. Many trials to date have compared one or more modalities against a medical control with variable results. To date, no single modality has been shown conclusively to be superior to others. As such, in this study we have not confined the endoscopist to one modality of treatment but allowed him to customize the method of endoscopic haemostasis according to the configuration, accessibility and rate of bleeding in any particular patient. Seventy-three patients with non-variceal upper gastrointestinal (GI) bleeding were admitted to the National University Hospital in Singapore between May 1, 1988 and April 30, 1989. All were gastroscoped and 48 were found to have chronic peptic ulcer. Twenty-nine (60%) with actively bleeding peptic ulcer or stigmata of recent haemorrhage (SRH) were treated endoscopically. Initial haemostasis was achieved in 27 (93%) patients. Seven patients rebled (26%) of which four underwent repeat endoscopic treatment. Of these four patients only one rebled again and required surgery. Permanent haemostasis was achieved in 23 of 29 patients (79%). The multimodality approach for the treatment of bleeding peptic ulcers gives the endoscopist flexibility in deciding on the best way to deal with a bleeding gastric or duodenal ulcer. Each instrument has its strengths and weaknesses and the right choice of instrument is often a critical factor especially in treating a bleeding ulcer in a situation where access poses a problem.