FAQ
The questions worth answering straight.
No marketing dodges. If the honest answer is "not yet" or "no", that's what's written below.
Answers
Nine questions, answered straight
Does my writing leave my computer?+
No. Essays, drafts and your publishing credentials stay on your own machine, and credentials are encrypted at rest. Nothing about how the app runs day to day depends on a server you don't control.
Do I need an API key?+
Not a separate metered one. Generation runs through the local claude CLI, the command-line version of Claude Code, instead of a pay-per-word API. That means there's no per-word bill stacked on top of the software itself.
What does it cost to run?+
Self-hosted has no software cost. You're paying for your own machine, its electricity, and whatever access to Claude you already have, nothing else. The full breakdown, including the two tiers that don't exist yet, is on the pricing page.
What happens if the AI is wrong?+
You see every draft before anything publishes. Separately, deterministic preflight checks (fixed, rule-based checks, not another AI guessing) catch structural problems: a post over a platform's character limit, a required field left blank, a reel script that won't fit its target length. Those get flagged before you approve a batch, not after it's live.
Can I edit before it posts?+
Yes. Drafts are generated for you to review and edit, and approval happens before publishing, not automatically after generation.
Which platforms does it work with?+
It generates content for LinkedIn, X/Twitter, Instagram, Threads and Bluesky. Scheduled auto-publishing covers four of those today: Bluesky, X, Threads and LinkedIn. Instagram's caption is generated and ready, but you post it yourself, since auto-publishing isn't wired up for Instagram.
Does it work offline?+
Your data being local and the app needing internet access are two different things. Essays, drafts and credentials stay on your machine regardless of connectivity. But generating the actual text calls out to Claude, so that step needs a connection. Voice narration and reel rendering, by contrast, run locally on your own GPU without a network round-trip.
What hardware do I need for voice cloning?+
A GPU on the machine running the studio, since narration is generated locally on it. Exact minimum specs aren't published on this site yet. If you're not sure your hardware is enough, ask on the contact page rather than guessing.
What if I don't use Substack?+
Substack is one way in, not the only one. You can paste an essay in directly, or import it from an Obsidian vault instead.
Still have a question?
This list is honest but not exhaustive. Ask directly and you'll get an actual answer, not a form-letter one.