Providing quality peer feedback and processing and using the received feedback are cognitively challenging activities of formative peer assessment (FPA) for students. The use of technological tools involving theoretically-grounded instructional scaffolds can be one promising solution to support these activities. Accordingly, this mixed methods research utilized scaffolds in an online environment that can activate self-regulated and co-regulated learning during FPA and investigated their impact on peer feedback provision and uptake. It included 70 ninth-grade students assigned into two groups to perform FPA activities for two writing tasks either with online regulation scaffolds (ORS) or without them. Using a new online FPA environment developed by the authors, they composed a story and poem in the first and second writing tasks, respectively. Results regarding peer feedback provision revealed that feedback in the group using ORS was briefer in two writing tasks, but it had significantly higher quality in the first writing task. In terms of feedback uptake, ORS led to more likelihood of reading peer feedback in the first writing task, but the percentage of students revising their performance was similar independent of ORS. Students’ experiences highlighted the role of feedback requests and goal setting and planning. Besides ORS, qualitative results called attention to the influence of task, learner, and feedback characteristics. Overall, this study concludes that ORS can be beneficial for improving peer feedback provision and uptake, especially in the first FPA practices for complex performance tasks. Moreover, it contributes to the development of the FPA model and guides the instructional design of online FPA environments with empirical evidence. • Online regulation scaffolds (ORS) were designed for peer assessment of writings. • ORS promoted peer feedback quality and reading peer feedback in the first writing. • Students emphasized the role of feedback requests and goal setting and planning. • Feedback use and task, learner, and feedback characteristics need more attention. • An assessment model for regulated learning was modified to remark task conditions.