Hi all -- I wanted to share a tool I've been developing that may be of interest to scholars who run online chat studies. Dyadic (
https://www.chatdyadic.com/) is a research platform designed for dyadic and small-group online conversations, supporting both human-to-human and human-to-AI interaction paradigms.
Here is what Dyadic currently supports:
-
Creating experimental and observational studies with configurable conditions (chats with text, chats with audio)
-
Deploying AI confederates alongside (or in place of) human participants, and deployment of AI suggestions in human-human chats to keep a conversation alive; API integrations with OpenAI, Anthropic, Gemini, and Hugging Face
-
Timed sessions, survey delivery before/during/after chat for in situ subjective ratings, plus real-time monitoring for researchers
-
Integration with Qualtrics, Prolific, SONA, etc.
-
CSV export of chat logs and survey responses, and many more features...
I'm particularly excited about the survey delivery feature, which allows researchers to send questions (Likert-type scales, feeling thermometers, open text boxes) to research participants, pairing their language data with self-reports.
Dyadic is in beta testing now, but it is effectively ready for deployment, and I have different projects using it right now. A preprint outlining its features and specifics is here:
https://arxiv.org/abs/2603.22227Please let me know what you think as you use Dyadic, and I'm always open to suggestions of how to make the tool better. Thanks for your time and interest!
--
Dr. David M. Markowitz
Associate Professor | Department of Communication
Michigan State University