FAQ
What unwrite writes, what a credit buys, who owns the work, the model behind it, how your data is kept private, and how the editorial review holds up.
Yes. The sample above is real output, and the free tier gives you one complete piece written by the same flagship model paying customers use. Read the actual writing, run the audit, export it, then decide. The free version is the real product, not a watered-down demo.
Non-fiction books/ebooks, white papers, and reports; long-form articles and essays; blog posts; email sequences; and conversion web copy. unwrite writes non-fiction, and it tells you before a fiction-shaped brief spends a credit. Each format is handled on its own terms, not one generic template: books/ebooks get chapters plus, on Pro, a Continuity read that re-reads every word and flags where a term, a promise, or the voice drifts; reports get an executive summary written last; web copy is built around page type and funnel stage.
Absolutely! Add your existing draft (from short-form content to an entire multi-chapter book), and unwrite loads it into the editor and runs it through the proprietary Editorial Engine every piece gets, in your voice, marking what reads as AI and what to tighten, so you fix in place rather than rewrite it out from under yourself. Built for anyone who works from existing drafts: a ghostwriter polishing a client's pages, a PR specialist tightening a release, a marketer sharpening a post. Add your own facts, background, or a voice sample alongside it, and only the sources you provide are ever cited.
A piece is one document on one brief, from first draft to finished. A credit is what creating it spends from your monthly balance: short-form costs 1, web copy 2 (every page arrives as two versions to compare), a narrative article 2, a white paper or report 3, a book/ebook 5, because long-form is drafted section by section. Those are the base costs. A piece run well past its usual length, or one where unwrite researches extra sources for you, costs more, and setup shows you what it adds before you generate. Revisions are included: 3 passes a piece on Free, 10 on paid plans. A new document or topic spends a new credit.
It is built not to. unwrite cites from sources it can verify, whether you provide them or it researches and checks them, and flags what it can't support instead of inventing it. No sources of your own? It researches them, checks each citation against the source it came from, and shows you what it found, rated for authority, to approve. The Source check covers up to 200 claims a piece, and tells you when a draft holds more. The facts stay grounded.
You do, completely: your voice samples, your sources, and every word generated. We use your material only to produce your drafts, and we never train models on your work.
Claude from Anthropic powers unwrite, with models chosen per task and wrapped in our proprietary Editorial Engine.
No. unwrite has no model-training pipeline of its own, and it never trains or fine-tunes any model on your writing. The model behind unwrite is Anthropic's Claude, used through Anthropic's commercial API under terms that forbid training on your content: Anthropic's Commercial Terms state that Anthropic may not train models on customer content from the services. Your material is used only to produce your drafts and run the features you invoke.
It is private to your account. What you bring (sources, briefs, voice samples, and drafts) is used only to help you, and is never pooled into a shared model or a shared dataset across customers. One customer's material never informs another customer's output. On a Teams plan, material is shared only inside the one organization that owns the workspace, and only with the roles that organization grants.
Your account and saved work live in a managed Postgres database in a United States region, encrypted at rest, and every connection is served over HTTPS with HSTS. Passwords are stored only as a salted argon2id hash, and sessions live in a secure, HTTP-only cookie. unwrite runs on cloud providers that maintain their own SOC 2 and ISO 27001 certifications; unwrite does not yet hold its own SOC 2 or ISO 27001 certification. The full statement is at unwrite.ai/security.
Because a chatbot hands you a draft and a shrug. unwrite hands you a draft in your own voice, with your claims checked against real sources, AI prose stripped, and an editorial audit that tells you what to fix before it ships under your name. Pro adds an independent Second read, and on books a Continuity read that re-reads every word and marks where a term, a promise, or the voice drifts.
Yes. Billing is monthly or annual, and canceling is self-serve from your account. You keep your paid features through the end of the period you've paid for; after that your account drops to the free tier. Your work stays put either way: every piece and its version history remains in your account to open and export, your saved voices stay in your account and follow you to any device you sign in from, and any top-up credits you bought never expire. Only the monthly plan credits stop refreshing.