MaiChat: A Text-based Dialogue Corpus Rich In Conversational Features

MaiChat is an English corpus of typed instant-messaging dialogues designed to capture spontaneous, interaction-rich written conversation. The dataset contains 42 dyadic conversations (84 native English speakers) collected via a custom-built real-time chat platform that records not only message content but also detailed behavioral and temporal metadata. In addition to full message logs, MaiChat includes fine-grained keystroke-level information such as inter-keystroke intervals, deletions, insertions, typing delays, turn timing, typing indicators, and screen activity logs. The corpus is particularly rich in conversational phenomena commonly observed in spontaneous interaction, including onomatopoeic and lexical laughter (e.g., “haha”, “lmao”), emojis and emoticons, filled pauses, expressive respellings (e.g., “yesssss”, “mmaybe”), intentional capitalization, and self-corrections. Conversations were collected between participants who know each other well, encouraging natural and informal exchanges. Designed with a transparent data collection protocol, MaiChat enables scalable collection while ensuring privacy and consent. The dataset supports research in conversational AI, cognitive modelling, human–computer interaction, behavioral biometrics, and the study of spontaneous written communication.