We sought to demonstrate the feasibility of an endoscopic approach to transapical aortic valve implantation (AVI), avoiding the morbidity of a thoracotomy incision. Using an experimental pig model, we performed three different approaches to transapical AVI, using a standard minithoracotomy (n=4), a robotic approach using the da Vinci telemanipulator (n=4) and an endoscopic approach using a port and camera access (n=4). The feasibility of the different techniques, exposure of the left ventricular apex, postoperative blood loss and total operative time were evaluated. Left ventricular apical exposure, 'purse-string' suture control and 33-F introducer access were successfully performed and confirmed videoscopically, fluoroscopically and at a post mortem in all 12 animals. The haemodynamics were stable in all animals. Mean intraoperative and postoperative (two-hour) blood losses were 88 and 65 ml with minithoracotomy, and 228 and 138 ml with the robotic and 130 and 43 ml with the endoscopic technique (P=0.26, P=0.14, respectively). There was no significant change in perioperative haematocrit (P=0.53). The mean total operative times were 1.4, 3.9 and 1.1 h (P=0.06), respectively. Percutaneous endoscopic and robotic transapical AVI are both feasible and can be performed in a timely manner with reasonable perioperative blood loss. Future research will focus on identifying optimal candidates for surgery based upon preoperative thoracic imaging.