Two people typing in the same paragraph is one of the genuinely hard problems in software. Agencies quote $60,000 to $250,000. We build a working core in one week for $5,000.
Not the screens. These are the parts that decide whether a real-time collaborative editor works or falls over:
Concurrent editing. You need conflict-free replicated data types or operational transforms. Last-write-wins destroys text in front of the user.
Presence and cursors. Everyone sees where everyone else is, updated continuously, without flooding the connection.
Version history. Every change retained so any earlier state is recoverable, without storing a full copy each keystroke.
Offline editing. Edits made with no connection have to merge cleanly on reconnect rather than clobber the document.
Permissions. View, comment and edit per document, plus link sharing that can be revoked.
The specific decisions that make a real-time collaborative editor hold up, and that the obvious version gets wrong:
A convergent data model. Edits are modelled so concurrent changes merge deterministically, using conflict-free replicated data types or operational transforms. Sending whole documents with last-write-wins deletes other people work in front of them.
The document is an ordered log of operations. Changes are appended as operations rather than overwriting a blob. Version history, restore and audit fall out of that structure for free.
Presence on its own channel. Cursors and selections are high-frequency and disposable. They travel separately from edits and are never persisted, so a room full of moving cursors does not bloat the document.
Offline edits replay on reconnect. Operations made with no connection are buffered and replayed against current state. This is precisely the case the convergent model exists to handle, rather than a special path bolted on later.
Compaction. Operation logs grow without bound, so they are periodically compacted into snapshots to keep storage and document load times under control.
Permissions enforced on the connection. Access is checked when the socket is established and on every operation, not only when the document first loads.
The gap
A studio quote for a real-time collaborative editor usually lands at $60,000 to $250,000 over five to twelve months. We build a launchable core in one week for a flat $5,000, deployed on your own accounts, with every line of code yours to keep.
A working product you can put in front of real users:
A rich-text editor with real-time collaborative editing
Live cursors, selections and presence
Comments and suggestions
Version history with restore
Permissions and revocable link sharing
Offline edits that merge on reconnect
And what we would leave out. Being straight about this is the only way a fixed price means anything:
Spreadsheets and presentations
Microsoft Office format fidelity, which is a large project on its own
Add-on and extension ecosystems
Enterprise compliance, retention and eDiscovery
You are getting the core of the same problem the big product solved, scoped to something launchable. Not a clone of the finished article.
Usually CRDTs, because they handle offline and reconnection more gracefully. We confirm the choice during scoping against how your users actually work.
The editing model can. The practical limit is how much presence traffic you want to pay for, and we size that with you.
A free 20 minute call. We will tell you if an app like Google Docs is a fit for one week, and what it takes.
Start here: what it costs to build an app in 2026, or how the week works.