This scientific opinion proposes toolboxes of welfare indicators for developing monitoring procedures at slaughterhouses for bovines stunned with penetrative captive bolt or slaughtered without stunning. In particular, the opinion proposes welfare indicators together with their corresponding outcomes of consciousness, unconsciousness or death. In the case of slaughter with captive bolt stunning, the opinion proposes a toolbox of indicators and the outcomes to be used to assess consciousness in bovine animals at three key stages of monitoring: (a) after stunning and during shackling and hoisting; (b) during neck cutting or sticking; and (c) during bleeding. For slaughter of bovines without stunning, a set of indicators and outcomes are proposed in another toolbox to be used for (a) assessing unconsciousness, before releasing bovines from restraint; and (b) confirming death before carcass dressing begins. Various activities—including a systematic literature review, an online survey and stakeholders’ and hearing experts’ meetings—were conducted to gather information about the specificity, sensitivity and feasibility of the indicators that can be included in the toolboxes. The frequency of checking differs according to the role of each person responsible for ensuring animal welfare. Personnel performing stunning, shackling, hoisting and/or bleeding will have to check all the animals and confirm that they are not conscious following stunning or before release from the restraint. For the animal welfare officer, who has the overall responsibility for animal welfare, a mathematical model for the sampling protocols is proposed, giving some allowance to set the sample size of animals that he/she needs to check at a given throughput rate (total number of animals slaughtered in the slaughterhouse) and tolerance level (number of potential failures). Finally, different risk factors and scenarios are proposed to define a ‘normal’ or a ‘reinforced’ monitoring protocol, according to the needs of the slaughterhouse.