Reuse profiles
One place for voice, banned words, product spelling and sentence length. Applied to every output automatically.
Features
Built for the work that happens after publishing
Every feature below exists because a content team told us which hour of their week they wanted back. Nothing here is a demo of what the technology can do — it is a list of jobs it removes.
Core
Source-aware drafting
Paste a finished article, a webinar transcript or a support document. Every derivative is built from that source, so the claims, the numbers and the structure carry across instead of being reinvented.
When the source changes, mark the piece for refresh and the derivatives that depend on it are flagged too. You will know which social post is now describing a feature you removed.
- Articles
- Transcripts
- Documentation
- Interview notes
What the engine reads from a source
- The argument and the order it is made in
- Numbers, dates and named claims
- Terminology you use consistently
- The action the piece is asking for
- Sections that work as standalone assets
The rest of the toolkit
Nine features, grouped by the part of the week they affect.
Approved claims
Store the figures and statements your team has cleared. Drafts reference that list rather than inventing numbers.
38 output formats
Each format carries its own structure, length target and tone adjustment. A thread is written as a thread.
Eleven languages
Produce the same brief in another language while keeping the approved claims and the call to action intact.
Shared library
Status, owner, word count and review date on every piece. Filter by channel or search by title.
Reuse scheduling
Set a date for a piece to come back around. The library surfaces it before the content goes stale.
Usage reporting
See what was produced, by whom, in which format, and how much of the monthly allowance is left.
Open exports
Clipboard, .txt, Markdown and HTML. Your library can leave whenever you want it to.
Client workspaces
On the Agency plan, each client gets its own profile, library and history under one login.
Under the hood
How the generator is wired
Two modes, one interface. Which one you use depends on whether you want to connect a model provider.
Browser engine
The generator on this site assembles a structured draft entirely in your browser using the options you choose. Nothing is uploaded, no key is needed, and it works offline once the page has loaded.
Best for outlines, structure, snippet sets and getting a piece started.
Connected provider
Point the tool at a small endpoint on your own server. That endpoint holds your provider key, calls the model, and returns finished prose. The key never reaches the browser.
Best for full drafts, translation and anything that needs real language generation.
Setup for mode 2 is documented in README.md and in the comments at the bottom of assets/js/generator.js.
See it work on your own content
The generator is open to everyone, with no account. Bring a paragraph from something you published and watch what it turns into.
- No account required to try it
- Drafts stay in your browser
- Download or copy anything you make