Writing and content decision
Lex
A competent developer can implement a useful subset (editor + LLM commands + sharing) but reproducing Lex's polished real-time collaboration, enterprise model integrations, privacy contracts, and product polish is multi-week work and retains nontrivial operational and trust costs that make keeping the paid product reasonable for many teams.
Visit website↗$12/mo
$144/yr
Not verified against a pricing page.
$100one-off80 h to build
$50/mo8 h/mo upkeep
On cash alone, building overtakes the subscription at 5 seats.
Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All Lex alternatives, with the arithmetic →
What a replacement has to do
- Create and edit rich-text documents collaboratively while invoking AI-assisted editing (rewrite/feedback/title generation) on selections or the whole document.
What it still won’t have
- Access to enterprise-model contracts and privacy protections Lex claims via their vendor relationships
- Priority email support / vendor-run support channel
- Polished cross-device UX (early-access iOS app / mobile polish)
- Any proprietary hosting, backups, and data retention policies Lex provides
What remains hard
- Brand trust
Loved by over 300,000 writers
- Compliance and regulation
All of your data on Lex is privately and securely hosted on our own database.
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 5 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a self-hosted web app using Next.js (React) + Yjs for CRDT collaboration, a Node/Express (or serverless) API with Postgres for persistent storage, and integrate OpenAI/Anthropic/Together APIs for LLM-powered edit commands. In scope: rich-text editor (TipTap or Slate) with selection-based AI commands (rewrite, feedback, title generation), real-time presence and sync via Yjs + WebSocket server, backend endpoints to save/load documents and generate shareable read-only/edit links, simple comments/versions UI, basic auth (magic links), and server-side usage tracking/billing telemetry. Out of scope: building mobile-native apps, enterprise contract negotiation, advanced moderation/compliance workflows. Include error handling for failed LLM calls, retries, and rate-limit responses, and provide unit tests for editor commands, API mocks for LLM calls, and an integration test for live collaboration syncing.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 4 cited sources+3
- Hard moats found in the evidence-3
- Evidence score57
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productLex — product
- official pricingLex Pricing
- open sourceONLYOFFICE/DocumentServer
- open sourceether/etherpad
Integrity checks
What held up, and what did not.
