RECAP: An End-to-End Platform for Capturing, Replaying, and Analyzing AI-Assisted Programming Interactions

2026-05-01

AI Summary

This work addresses the lack of fine-grained recording and analysis capabilities for the full interaction process between developers and AI programming assistants, which has hindered the understanding of authentic collaborative behaviors. We propose an end-to-end platform built upon the VS Code extension framework that unobtrusively captures AI dialogue exchanges and code editing events, fusing them into a unified timeline to enable context-aware replay. The platform incorporates an extensible analysis layer integrating behavioral classification models and AI-dependence metrics. For the first time, it enables fine-grained, multidimensional tracking of the entire AI-assisted programming interaction. Deployed in a university software engineering course, the system successfully collected 2,034 prompts and 8,239 code edits from 41 students, establishing a high-quality dataset and analytical foundation for studying developer–AI collaboration patterns.

Abstract

Understanding how developers interact with AI coding assistants requires more than chat logs or git histories in isolation; it requires reconstructing the full context: which prompt led to which edit, what the developer tried and discarded, and how their strategy evolved over time. We present RECAP (Replay and Examine Captured AI Programming), an open-source platform that (1) passively records AI chat sessions and fine-grained code edits inside VS Code without disrupting the developer's workflow, (2) merges them into a unified timeline for interactive session replay, and (3) exposes an extensible analysis layer, with example modules for behavioral classification and AI reliance measurement. Deployed in a university software engineering course, RECAP captured 2,034 prompts and 8,239 code edits from 41 students across a multi-week project. We demonstrate how the platform's linked data and replay capabilities enable analyses of developer-AI interaction patterns that no single data source could support.

Problem

• AI-assisted programming
• developer-AI interaction
• interaction context
• programming behavior
• code editing

Innovation

• AI-assisted programming
• interaction replay
• fine-grained code editing
• behavioral analysis
• developer-AI collaboration