Overview
What Enrai is, where it fits, and how the workspace turns coding work into an operational flow.
Stop losing engineering flow between editor, terminal, preview, and AI.
Enrai turns local software work into a single operational surface. Instead of jumping between tabs, shells, scripts, and partial agent state, your team works in one workspace with files, PTY-backed terminals, previews, sessions, rollouts, and queued Codex work.
The problem
Most “AI coding” setups still fracture the actual work:
- your editor knows the code but not the shell state
- your terminal runs the process but not the rollout
- your preview shows runtime output but not the task queue
- your AI chat responds, then loses continuity when work moves elsewhere
Code changes, shell output, previews, and AI instructions live in different tools with weak continuity.
Ad-hoc scripts and manual terminal workflows do not give you resumable tasks or reliable handoff points.
Without sessions, workdir control, and queue state, automation becomes opaque and brittle.
How Enrai works
Enrai mounts a real directory and keeps files, explorer state, preview paths, and PTY terminals pointed at that workspace.
Use the browser editor, terminals, preview tabs, and Enrai chat without losing shell state or file locality.
Codex rollouts, queued tasks, and session metadata give you a workflow that survives longer than a single prompt.
What the workspace includes
- Monaco-based code editing with live file operations.
- PTY-backed terminals with shell continuity.
- Local preview tabs for web apps and runtime checks.
- File explorer and search aligned to the active workspace root.
- Enrai chat with session, rollout, and task controls.
- Portable and split deployment modes depending on how you operate the stack.
Example output
Instead of “here is a patch” with no execution context, Enrai preserves the workdir, terminal state, session path, rollout history, and queue status needed to continue, inspect, or recover the task later.
Value by role
Stay in one workspace while editing, running, previewing, and coordinating with Enrai.
Deploy a controlled workspace surface for trusted environments instead of stitching together ad-hoc local tooling.
Queue work, inspect status, resume rollouts, and integrate Enrai into larger callback-driven flows.
Repository layout
frontend/ Main editor application
backend/ Express API, WebSocket server, PTY manager, Codex integration
documentation/ Fumadocs-powered documentation site
bin/ CLI entrypoints
scripts/ Bootstrap and utility scriptsStart here
Use the shortest path to launch Enrai locally, in Docker, or in portable mode.
Open Getting StartedChoose between portable and split hosting based on trust boundary and ops needs.
Open DeploymentUse the task queue and rollout continuation flow for longer-running agent operations.
Open Task Queue