Employing multimodal computer-mediated communication (CMC) for online language learning and teaching has gained momentum worldwide due to the emergence of various digital modes, such as text, image, audio, and video, for online communication. This pilot study aimed to explore students' learning experiences with multimodal CMC tasks through Instagram. Thirty first-year students at an Indonesian university completed seven CMC tasks, consisting of information gap, reasoning gap, and opinion gap tasks, through three Instagram communication channels: text chat, voice chat, and video chat. Pre- and post-study surveys, journal reflections, and interviews were analyzed using a mixed methods approach. Findings revealed that students overall positively perceived their experiences with tasks delivered through Instagram video, audio, and text chats. They also reported that paralinguistic features afforded by the multimodal Instagram channels-- such as emojis, GIFs, images in text chat, intonation in voice chat, and gestures in video chat-facilitated effective communication. However, challenges such as poor internet connections, lack of consciousness of student agency when interacting in video chats, and high anxiety at the beginning of task implementation were also documented during student task performance. The findings suggest that the use of multimodal CMC channels affords greater accessibility and provides multimodal affordances for language learners to communicate using rich semiotic resources. They can strategically draw upon their digital literacy skills to convey messages during meaningful task interaction. Nevertheless, language instructors should consider the availability of internet infrastructure and students' language proficiency prior to utilizing multimodal CMC channels as language learning tools.