Abstract Background Flexible endoscopy plays an important role in digestive health. However, access to endoscopy is limited in many rural areas throughout the world. Training non-physician personal to perform diagnostic endoscopy and to transmit images to a central hospital, where experienced endoscopists can review the procedures, may improve digestive health for patients in remote areas. The aim of this study was to evaluate the diagnostic quality and accuracy of upper-GI tele-endoscopy. Methods Fifty patients scheduled for EGD underwent upper-GI tele-endoscopy. The procedures were observed simultaneously by the endoscopist and a gastroenterologist observing from a remote station connected by 4 integrated services digital network telephone lines. The interpretation of the findings by both were compared and concordance for diagnosis of major and minor lesions was analyzed. Results Tele-endoscopic image quality was adequate to support diagnosis of abnormal lesions by the remote observer. Technical issues included worsening image quality caused by mild pixelation during rapid endoscope movement and rare loss of the telephone lines. The endoscopist identified 47 different major and 44 minor findings in the 50 patients. The observer missed one major lesion (columnar-lined esophagus) because of suspected inflammation and described 10 non-existing major lesions (sensitivity 98%: 95% CI[89%, 99%], specificity 80%: 95% CI[66%, 90%]). Some of the differences were because of interobserver variability. Conclusions Upper-GI tele-endoscopy by using telephone lines has good diagnostic quality and is highly sensitive with regard to major findings. The misinterpretation of certain findings (esophageal ring, gastric erosions) may be caused by interobserver variability. The data strongly suggest that endoscopist and observer see similar endoscopic views.