<![CDATA[Purpose: This study examines how fifth- and sixth-semester Japanese as a Foreign Language (JFL) students at Universitas Negeri Manado (UNIMA), an outer-island public university in North Sulawesi, Indonesia, use digital technologies to support self-regulated learning (SRL). It explores students' technology-use patterns, the SRL dimensions these tools support, and how local contextual factors shape technology-enhanced language learning. Methods: Using a qualitative case study design within an interpretive constructivist framework, data were collected over eight weeks from 30 participants through semi-structured interviews, an adapted Motivated Strategies for Learning Questionnaire (MSLQ, α = .87), classroom observation, and document analysis. Data were analyzed using thematic analysis alongside descriptive statistics. Findings: Participants showed a moderate overall SRL profile (M = 3.55), with goal-setting as the strongest dimension (M = 3.82) and self-evaluation as the weakest (M = 3.29). Four themes emerged: (1) tool selection driven by cost and peer recommendation, dominated by YouTube (93%), Duolingo (90%), and Anki (80%); (2) uneven digital support across SRL phases, with stronger support for practice than for self-reflection; (3) learning strategies adapted to limited local digital infrastructure; and (4) peer-based self-regulation supported through WhatsApp study groups. Research Implications: These findings suggest a need for explicit teaching strategies that build students' self-reflection skills, greater investment in campus digital infrastructure, and formal recognition of peer study groups as part of language learning support. Originality: This study introduces the concept of phase-differential digital scaffolding uneven technological support across different stages of self-regulated learning and shows how students adapt their strategies within an under-resourced, outer-island setting, extending existing research that has mainly focused on urban, well-resourced contexts. Limitations: As a single-site case study with 30 participants, findings may not generalize to other institutions or regions; self-reported data may also be subject to social desirability bias.]]>
