Abstraction identification is an important requirements engineering (RE) task that can support elicitation and modeling by discovering significant domain terms. Prior work, notably AbstFinder and RAI (relevance-driven abstraction identification), has introduced the core ideas, but offered only limited tool support. In this paper, we present an abstraction identification tool, Co-AI, built on the Google Colab environment allowing the users to run the tool within their web browsers, promoting tool adoption and extension. Co-AI integrates the Wikipedia pages as the domain corpus, and identifies the candidate abstractions with a set of natural language processing (NLP) patterns. Co-AI is freely available at: https://colab.research.google.com/drive/1ur5KILoi_n-3KY0_vJcMBQDtiSYgcYeP?usp=sharing and we welcome the community’s feedback of our tool.