Can Coding Agents Solve Repository-Level Issues with Rendered Code? An Exploratory Study of Visual Representations
2026-08-10 • Human-Computer Interaction
Human-Computer InteractionArtificial Intelligence
AI summaryⓘ
The authors explored using pictures of code (rendered code) instead of plain text to help coding agents work with large codebases. They found that showing code as images can reduce the amount of text the agent needs to process, saving some computational resources, but these savings don’t always scale up well. While this visual approach keeps the same level of accuracy in fixing code, it doesn't improve beyond the limits of the agent technology and can sometimes be unstable if too compressed. The authors conclude that using visual code makes sense particularly when reading the raw code text is a big slowdown, but it helps less during trial-and-error patching.
rendered codevisual compressionagentic codingrepository-level repairprompt-token costpatch testingstatic code understandingcode representationagent architecture
Authors
Weijie Liang, Yuanfeng Song, Xing Chen, Caleb Chen Cao, Sirui Han, Yike Guo
Abstract
Visual modality has recently been explored as a way to compress textual tokens, including rendering code as images for static code understanding. We study whether this representation can serve as operational context for agentic coding, where an agent must navigate repositories, edit source files, and verify executable patches. Using SWE-bench Verified, we evaluate rendered code in repository-level repair workflows and introduce controlled agent settings to separate unguided repository exploration from more structured repair stages. Our results show a mixed picture. Rendered code consistently reduces prompt-token cost, but the savings do not increase linearly with the nominal visual compression ratio. It largely preserves end-to-end repair accuracy, but does not overcome the performance limits of the underlying model or agent architecture, and can become unstable under aggressive compression. Further analysis suggests that visual code is most useful when raw source reading is a major bottleneck; once repository localization is structured, much of the remaining cost comes from patch--test trial-and-error, where visual compression has limited leverage. Overall, our study positions rendered code as a viable but conditional compression mechanism for realistic coding agents.