Skip to Oxygen content

Why an Agent Studio Should Not Look Like VS Code

If you ship a code editor, people will ask why it is not VS Code. Fair question. Wrong product goal. Oxygen includes editors, a script runner, and a sandbox shell — then refuses to make them the hero. The hero is the agent loop that turns files into analysis, visualizations, and small apps.

The layout is a product decision

Chat sits in the center. One sidebar handles chat sessions and file/folder browsing. The other opens code and editors when you need to inspect or tweak. That is enough structure for serious work without cosplaying a full IDE.

Who this is for

  • Analysts who live in exports and need cited narratives or charts.
  • Operators who turn SOPs and CSVs into briefings.
  • Makers who want a UI or script out of a session — not a new primary editor for their career.

Power users still get a sandbox shell and package control. Depth is available; it is just not the default identity of the product.

What “done” looks like

  • A verified brief or memo grounded in the corpus.
  • A visualization or graph output from a flow.
  • A small UI from the front-end builder.
  • A scripted reshape of a CSV you can re-run next week as an agent.

FAQ

Can developers use it?
Yes — especially with workspace mounts and the sandbox — but Oxygen optimizes for agent-driven analysis and asset creation, not daily IDE ergonomics.
Is there a terminal?
Power users can use the sandbox shell. Most users stay in chat and flows.