One Editor per Project
Only one editor can have a project open for writing at a time. Two editors that write the same project overwrite each other's work, so the first editor to open a project owns it, and any further editor opens it read-only.
How ownership works
When an editor opens a project, it takes the project's lock: EditorSession/Owner.lock in the project folder. Next to it, Owner.info records who owns the project (user and computer, process, start time) and is refreshed regularly as a heartbeat. The editor writes a .gitignore and a .dvignore into the EditorSession folder, so version control does not pick it up.
The operating system releases the lock when the editor process ends, also when it crashes. A project is therefore never locked for good. When an editor finds a lock that a crashed editor left behind, it takes the project over and logs Took over the project from an editor that did not close it. The recovery dialog then offers that session's unsaved work.
Processes that the owning editor starts itself, such as the background asset processors, work in the project on behalf of the editor and don't need the lock.
Opening a project that is already open
If another editor owns the project, the editor asks what to do:

| Situation | Choices |
|---|---|
| The project is open in another editor. The message names the owner, its process, when it started and when it was last seen. | Open opens the project read-only. Retry checks again, for example after you closed the other editor. Close exits. |
| The editor that locked the project is no longer running, but the operating system still reports its lock file as in use, for example because of a virus scanner or a stuck process. | Yes takes the project over, and offers that session's autosaved changes next. Open opens it read-only. Close exits. |
| The project is on a network drive and may be open on another computer. Network shares don't always report who holds a lock, so the editor goes by the heartbeat: an owner counts as gone once it has not reported in for 45 seconds. | Open opens it read-only. Retry waits and checks again. Yes takes it over anyway, after a second confirmation. Close exits. |
Only take a project over when you are sure the other editor is not running. If both editors keep running, both write to the project and the later save wins.
Read-only mode
While a project is read-only, nothing in it is written. This covers every way of changing a project: saving, autosave, importing, background transforms, Asset Browser file operations, drag and drop, menu actions and shortcuts, and MCP tools. Actions that would write are greyed out, with the reason in their tooltip.
You can tell that a project is read-only by:
- the status bar of the main window, which starts with READ-ONLY: and the reason, for example The project is open read-only because it is being edited by jane on STUDIO-PC.
- a notice at the top of every document, The project is read-only.
When the project was opened read-only because another editor owns it, the document notice has a Try again button. If the other editor has closed the project in the meantime, the editor restarts with the project open for writing, so the documents load from the files as they are now. If the project is still in use, it tells you who has it.
A single document can also be read-only, for example a document that is opened for viewing only. Its notice reads Opened read-only and offers Edit anyway. That is unrelated to project ownership.
Losing the lock
If another editor takes the project over while you are working, for example after a network drive dropped out and the other user chose Yes, your editor stops writing to the project at once. It logs Another editor took the project over, stops background processing and shows This editor lost the project lock, so it no longer writes to the project in the status bar. Your unsaved changes stay in the open documents. To keep them, copy what you need, close the editor and reopen the project once the other editor is done with it.
Command-line options
| Option | Description |
|---|---|
-readOnly | Opens the project read-only without taking its lock. Nothing in the project is saved, transformed, imported or changed in any other way. Use it to look at a project while somebody else edits it. |
-projectOwner <pid> | Used by processes that the editor starts: the process id of the editor that owns the project. The process then works in the project without taking the lock. If that editor no longer owns the project, the process asks for the lock itself. You don't normally pass this by hand. |
Editors and tools that can't show a dialog, such as EditorProcessor and editors started with -nostartupdialogs, never guess: if the project is owned by another editor they log The project '...' is open in another editor and exit with return code 2, so a script that expects to write does not silently run read-only. Pass -readOnly if read-only access is what you want. Project checks run with EditorProcessor -checks only read, so they never need the lock.
Troubleshooting
| Symptom | Cause |
|---|---|
| The project always opens read-only, without asking | The editor could not create its lock file, for example because the project folder is write-protected. The log says The project's lock could not be created. |
| A build script fails with return code 2 | An editor has the project open. Close it, or run the script with -readOnly if it only reads. |
| The editor asks about another editor that you know is closed | On a network share, wait 45 seconds and press Retry, or take the project over with Yes. |