Weld repair and reclamation are often intensive operations with the possibility of exposing personnel to welding hazards for a prolonged time. Robotic welding has the potential to improve health and safety by removing the operator from the immediate area of the weld. Automation can also improve control of welding parameters in critical weld repair situations such as in service repair of pipelines. In spite of these obvious advantages it is often difficult to justify the use of robotic automation for non repetitive weld reclamation and repair situations due to the time consuming manual programming overhead. The present paper describes the development and evaluation of a technique for rapidly generating off-line robot programs for weld repair. An important step in this process is determination of the surface profile and the extent of the damage area. Whilst several standard profilometry systems are available; they are often very costly and it is often difficult to define the boundary of the repair area. The system described here uses a personal computer linked to a robot controller and GMAW welding system. A single camera and a projected grid pattern from a low power laser are used to obtain information about the profile and position of the object to be welded. Manual verification of the wear area is performed on the captured image. The program incorporates welding process algorithms which determine the appropriate parameters for the desired build up. The paper describes the technique and shows how the information obtained is automatically converted to a robot welding program which implements the weld deposition. Validation of the approach on a simulated pipeline repair is described.