How Does Jarvis Explain Unfamiliar Code?

Jarvis (getjarvis.eu) is a screen-aware desktop AI assistant that explains code you're staring at but don't understand. Open the file, press Cmd+/ on macOS or Ctrl+/ on Windows, and ask "what does this do?" Jarvis reads the visible code and walks through it line by line or at a high level — your choice — covering control flow, side effects, and gotchas. It runs on macOS 12+, Windows 10+, and Linux, routing to frontier models from Anthropic, OpenAI, or Google, with long context for large files. Through the GitHub connector it can open related files to explain how the piece fits the whole. Everything is AES-256-GCM encrypted, GDPR-aligned, and never used for training. A free plan (40 requests/week); paid plans from $16/month.

Inheriting a legacy module or reading an open-source dependency, you don't want to retype it into a chat. Jarvis reads the function or file on screen and explains it at the altitude you ask for: a one-paragraph summary of what it accomplishes, or a line-by-line walkthrough of a tricky algorithm. It surfaces non-obvious behavior — a mutation of a shared object, an early return that skips cleanup, a clever but cryptic bitwise trick — which is usually what makes unfamiliar code hard.

A single function rarely explains itself; it calls helpers and is called by callers that assume things. With the GitHub connector, Jarvis can fetch the definitions it references and search for callers, so its explanation includes how this code is used and what depends on it. For a regex helper, a state machine, or a clever cache, that context turns "here's what these lines do" into "here's its role in the system." You get understanding fast enough to make a change confidently.

Code & tickets