no signup · 3 free generations a month

Posts that sound like you — not like AI.

Paste rough notes or a news topic. Get three social posts with the robot tells stripped out — no tidy morals, no hype words, no invented facts. Ready for Threads, X, or LinkedIn. Try it in one paste, below.

Your material

My notes
Today's news

Drafts can only use what's here. Nothing gets invented to fill space. Paste and hit Generate — that's the whole thing.

Make it sound more like you optional — sharpens the voice a lot

Run more than one account? Save each voice as a profile and switch in one click. Stored in this browser.

The single biggest lever on how human it sounds. Real posts beat polished ones. Saved in your browser.

Drafts are editable — click into one and tweak it before you copy.

The only material the AI was allowed to use

How it works

Grounding first. Voice second. Nothing else.

01 · MATERIAL

You provide the facts

Rough notes about your day, or a keyword that pulls today's real headlines. This becomes the only source of truth for the draft.

02 · GROUNDED WRITING

The AI writes inside the fence — then gets audited

Every name, number, and event in the draft must exist in your material. Then a second, independent model pass audits each draft and flags anything it can't trace back to your material. You see the result on every draft.

03 · YOUR VOICE

It sounds like you

Tone is copied from example posts you provide, not from generic instructions. The usual AI tells — tidy morals, balanced sentences, hype words — are banned outright.

Why grounding

Most AI writing tools are confident liars.

While building the automation engine behind this tool, we asked an AI to write about "a problem you hit today" without telling it what happened that day. It produced a detailed, convincing debugging story — involving tools the project didn't even use. Completely fictional.

One invented post can sink an account people are supposed to trust. So this engine works inside one rule: if it's not in your material, it can't be in your draft. And you can always expand "the material the AI was allowed to use" under your results to check.

That transparency is the product. Post with your name on it, knowing where every word came from.

Pricing

Free to try. One payment if it earns its keep.

$0
  • 3 generations per month
  • Both modes — notes and news
  • No signup, no card
Try it now
$39 once · founder license
  • No monthly limit (200/mo fair-use cap)
  • Lifetime — pay once, keep it
  • Founder pricing, goes up later
Get a founder license — $39

Your license key is generated instantly at checkout — paste it into the box below the generator. Not useful? Reply to your receipt email for a refund.

Honest answers

What this is — and isn't.

Which platforms does it work with?

Any platform where you paste text: Threads, X, LinkedIn, Bluesky, Mastodon. Drafts target 350–500 characters — inside Threads' and X's limits, and a good length everywhere else. You copy, you post, you stay in control of what goes out under your name.

Why doesn't it post automatically?

Because reviewing what goes out under your name is a feature, not friction. If you do want full automation for Threads — daily schedule, self-review pass, zero clicks — that exists as a separate self-hosted kit: Threads Autopilot ($29). Same grounding engine, running unattended.

Is my writing stored?

Your voice profile and notes live in your browser's local storage, not on our servers. Notes are sent to the generation API for the seconds it takes to write your drafts, plus OpenAI processes them as the model provider. We keep an anonymous counter (how many generations happened) — not your content.

Will readers know it's AI?

The usual giveaways — invented specifics, tidy moral endings, uniform paragraphs — are exactly what this tool bans. But the honest answer is: the output is only as authentic as your material. Real notes in, posts that sound like your life. Empty hype in, hype out. That part stays on you.

What happens if the free tier gets abused or costs spike?

The free tier is 3 generations a month, and founder licenses carry a 200/month fair-use cap. Those limits are what let a one-person product keep a lifetime deal honest instead of quietly killing it later.