Background: The universally accepted best practice protocol for monitoring patients who receive intravenous thrombolysis for acute ischaemic stroke was established in the 1990s. However, the protocol is burdensome for nurses, disrupts the sleep of patients, and is potentially less relevant in patients at low risk of symptomatic intracerebral haemorrhage. We aimed to assess whether implementing a low-intensity monitoring protocol would be as safe and effective as standard high-intensity monitoring for patients with acute ischaemic stroke at low risk. Methods: OPTIMISTmain was an international, pragmatic, multicentre, stepped-wedge, cluster-randomised, controlled, non-inferiority, blinded-endpoint trial conducted at hospitals (clusters) in eight countries. It was designed to test the non-inferiority of a low-intensity monitoring protocol to a standard protocol among consecutive adults with acute ischaemic stroke who were clinically stable with mild to moderate neurological impairment (score <10 on the National Institutes of Health Stroke Scale) within 2 h of initiation of intravenous thrombolysis according to local guidelines. Participating hospitals were randomly allocated to three sequences of implementation across four periods, stratified by country and projected numbers of participants, in which sites switched from standard monitoring (control) to low-intensity monitoring (intervention) in a stepped manner. The low-intensity monitoring protocol included assessments of neurological and vital signs every 15 min for 2 h, every 2 h for 8 h (vs every 30 min for 6 h for standard monitoring), and every 4 h (vs every 1 h for standard monitoring) until 24 h after thrombolysis. The primary outcome was the proportion of participants with an unfavourable functional outcome defined by a score from 2 (indicating some disability) to 6 (death) on the modified Rankin Scale at 90 days, measured by research staff masked to group allocation. The non-inferiority margin was set at 1·15 for the risk ratio (RR) in the intention-to-treat population. A generalised linear mixed model was used for analysis with adjustments for cluster (hospital site) and time (6-month periods from April, 2021), and imputation of missing outcome data. This trial is registered at Clinicaltrials.gov (NCT03734640) and the Australian New Zealand Clinical Trial Registry (ACTRN 12619001556134p) and is completed. Findings: Of 181 hospitals assessed for eligibility, 120 hospitals agreed to join the trial and were randomly allocated between April 28, 2021, and Sept 30, 2024; however, one hospital withdrew, one was not activated, and four did not enrol any patients. Overall, 4922 participants were enrolled at 114 hospitals, with 2789 participants assigned to the low-intensity monitoring group and 2133 to the standard monitoring group. 809 (31·7%) of 2552 participants in the low-intensity group and 606 (30·9%) of 1963 in the standard monitoring group had a modified Rankin Scale score of 2–6 at 90 days (RR 1·03 [95% CI 0·92–1·15], pnon-inferiority=0·057). Symptomatic intracerebral haemorrhage occurred in five (0·2%) of 2783 patients in the low-intensity group and eight (0·4%) of 2122 patients in the standard monitoring group. The numbers of participants with a serious adverse event were similar between the low-intensity monitoring group (309 [11·1%] of 2789) and the standard monitoring group (240 [11·3%] of 2133). Interpretation: OPTIMISTmain provides weak evidence that low-intensity monitoring is non-inferior to standard monitoring in patients with a mild or moderate level of neurological impairment who receive thrombolysis treatment for acute ischaemic stroke. Hospitals could consider incorporating this approach into stroke services according to local circumstances. Funding: National Health and Medical Research Council of Australia; New South Wales Health Investigator Development Grant; University of New South Wales Medicine Non Communicable Diseases Theme Early–Mid Career Research Seed Grant Scheme; Medical Research Future Fund (for conduct in Australia); and Genentech (for conduct in the USA).