Özgür Kurucan

I built Argus after repeatedly getting frustrated with how opaque Claude Code sessions were there was no clear way to understand why a session cost more, looped, or behaved unexpectedly, so I designed a lightweight VSCode extension that parses the local .claude session logs and turns them into a visual, step-by-step time machine debugger, showing tool calls, token usage, retry loops, and file interactions in a way that s actually explorable; the core idea was to bring observability (something we take for granted in backend systems) into AI coding workflows, so I built it using a simple TypeScript + React stack with a focus on fast local analysis, heuristic-based insights, and zero external dependencies, iterating quickly by analyzing my own sessions and discovering patterns like repeated file reads and costly retry loops, which ultimately shaped the product into both a debugging and cost-visibility tool for developers using Claude Code.

Terminal or editor-first UI? How do you prefer to work with AI coding agents?

I've been using Claude Code daily for months, and while it's incredible, I kept running into the same frustration: I had no idea what was actually happening during sessions. Why did that refactoring cost $12? Why did it take 15 minutes? Is Claude reading the same config file over and over? The answers were technically there buried in .jsonl log files that I'd never realistically parse by hand. So I built Argus, a VSCode extension that just... reads those files and makes sense of them