ORGN Studio

ORGN Studio is the browser application for planning, orchestrating, and reviewing AI-assisted engineering work. Open it at cde.orgn.com.

It is not a web version of ORGN CDE — the desktop IDE. If you are looking for a VS Code fork with local folders, extensions, and SSH attach to cloud worktrees, start with CDE.

Why does Studio exist alongside CDE?

ORGN ships two developer surfaces that share the same identity, projects, and sandboxes — but solve different jobs.

Studio is closer to an engineer-focused task system (structured work tracking, assignees, milestones, agent assignment) with a browser workspace for chat, files, and terminal — not a duplicate desktop IDE in a tab.

Both products connect to the same projects and worktrees. A task you create in Studio can be executed in a browser workspace or picked up later in CDE on the same branch and sandbox.

When to use which

Rule of thumb: use ORGN Studio to set up the project, structure the work, and collaborate in the browser. Use ORGN CDE when you want a native IDE for sustained editing on the same worktree.

See also the CDE-side comparison on CDE overview.

What ORGN Studio is

Studio is three layers, not one monolithic "web IDE":

1. Project and task management

A project is one Git repository inside your team. Each project has:

• A task board — priorities, assignees, statuses, labels (structured work tracking, not just chat history)
• A dashboard — open worktrees, pull requests, and what needs attention
• Settings — models, secrets, sandboxes, usage

This is the coordination layer. Think mission control, not an editor replacement.

2. Browser workspace

When you open a task or start an agent session, Studio opens a browser workspace — chat, file tree, terminal, and diff review in the browser. Internally this is Code Mode; do not confuse it with ORGN CDE, which is the desktop product name.

Work runs inside TDX sandboxes — the same confidential compute ORGN CDE attaches to over SSH.

3. Agent orchestration

Studio lets you:

• Run interactive sessions — describe work in chat; the agent reads, edits, and runs commands
• Assign autonomous agents to tasks — agents spin up worktrees and execute without you in the loop
• Generate feature specs before code — structured ideation that becomes tasks

Agent activity, inference routes, and sandbox metadata are reviewable for security workflows. See Platform trust and ORGN Scanner.

How work is structured

Team
 └── Project (one Git repository)
      ├── TDX Sandbox (confidential runtime — shared by all worktrees)
      ├── Tasks (project backlog)
      ├── Feature specs & context docs
      ├── Settings (models, secrets, usage)
      └── Worktrees (isolated Git branches inside the sandbox)
           └── Sessions (agent chats at /chat/:id)

Execution flow: pick a task → open or create a worktree inside the project sandbox → work in sessions (browser workspace). Feature specs and settings sit alongside the backlog at the project level — they are not inside worktrees.

For panel-level detail, see Workspace view and Tasks.

Confidentiality

Studio workspaces and agent tool calls run inside Intel TDX Trust Domains. Inference routes through ORGN Gateway — pick TEE models for hardware-attested inference or ZDR models for policy-based retention.

Confidentiality is verifiable — not a vendor promise. See Platform trust.

Documentation

Related

• Gateway quickstart — API keys and model catalog outside Studio
• Code security — vulnerability scanning (distinct from ORGN Scanner attestation)
• Platform trust — canonical confidentiality reference