FAQ
Questions, answered.
Can't find what you're looking for? Send us a note.
Product
Does it really sound like me?+
The quality is directly proportional to the writing profile you set up. Three real sample posts plus a few banned phrases gets you 80% of the way. Tighten your tone, audience, and structure notes over time and the voice gets sharper with every generation.
Where does the research come from?+
We use semantic web search via Exa to pull recent articles, studies, and news relevant to your topic. Every draft ships with citations so you can verify before posting. Results are cached for 24 hours per query, so repeat topics stay snappy and cheap.
How are the three variants different?+
Each generation produces three angles: a hook + story (vivid moment, lands on a lesson), a data-driven take (anchored on a specific number or pattern from research), and a contrarian one (challenges a common assumption). Same voice, different framing — pick the one that lands or remix across all three.
Will it post to LinkedIn automatically?+
No. We're a generator, not a publisher. Automated posting violates LinkedIn's ToS, and you should read every post before publishing anyway. Copy-paste takes two seconds and the output is LinkedIn-friendly out of the box (clean line breaks, no markdown).
Can I use this for Twitter / X or Threads?+
LinkedIn-first today. The writing profile system maps cleanly to other platforms — multi-platform support is on the roadmap.
What AI model powers it?+
Claude Sonnet 4.5 by Anthropic. It's meaningfully better at writing in a specific voice than general-purpose models, and we use prompt caching on your writing profile to keep generation costs reasonable.
Account & billing
How does the free plan work?+
50 generations per month while we're in private beta. No credit card required. Each generation produces three variants — that's effectively 150 voice-matched drafts a month for free.
What's included on Pro?+
500 generations per month, priority support, and lifetime beta pricing if you join while we're in private beta.
Can I cancel anytime?+
Yes. Self-serve cancel through Stripe's customer portal from your billing page. You keep access until the end of the billing period, and your post history stays exportable.
Do you offer refunds?+
Within 14 days of your first paid charge, yes — no questions asked. Email us and we'll process it.
How do I export my data?+
From the account page, click “Download JSON”. You get your full writing profile and every generation you've ever made.
How do I delete my account?+
From the account page, scroll to “Delete account”. It cascades — profile, generations, and connected sign-in providers are removed permanently.
Privacy & data
Do you train on my writing samples?+
No. Your sample posts and generations are not used to train any model. Anthropic, our model provider, does not train on API traffic by default. We use your samples only as few-shot context inside your own generations.
Who can see my generations?+
Only you. Generations are scoped to your user account. Our team can see aggregate metrics (token usage, generation count) for billing and support, but does not browse generation contents unless you explicitly ask us to in a support ticket.
Where is my data stored?+
On managed Postgres in the EU/US (depending on your region). All connections are encrypted in transit. Backups run daily and are retained for 30 days.
Are my passwords stored anywhere?+
No. We use NextAuth (Auth.js) with Google OAuth or email magic links — there are no passwords to leak.
Troubleshooting
A variant contains a banned phrase. What do I do?+
We flag it on the variant card before you copy it. Edit the line out before posting, or hit Generate again — the model is told never to use your banned phrases, but slip-ups happen with creative writing.
Generation is slow.+
First generation in a session takes ~10 seconds (we pull research and the writing profile is uncached). Subsequent generations within a few minutes are faster — the writing profile is served from Anthropic's prompt cache.
Research came back empty.+
Exa couldn't find relevant recent results for that topic. The model still writes — it just leans on your profile and structure rather than citations. Try a more specific topic or paste a link to anchor the research.
Hit my monthly limit early.+
Either upgrade to Pro for 10× more generations, or wait until the 1st of next month. You can see usage and reset date on your billing page.