Alpha — early access. This page describes an early-stage product; features and pricing may change.

Quillecho: a walk through what using it looks like

Three scripted scenarios. Every screen below is staged for this pitch — no real archive was ingested and no model generated these drafts.

Autoplaying — simulated: learning from the archive (step 1 of 3)

Illustrative example

Step 1 — Connect one archive source

A writer pastes in their Substack export URL. Quillecho pulls the published pieces it can reach and starts building a private voice model from them.

↓
Voice model built: short punchy opening lines, frequent rhetorical questions, a habit of ending sections with a one-line turn, informal-but-precise word choice. (Illustrative summary — an actual model would not surface as a plain-English list like this; shown here for the pitch, not as a real UI element.)
Illustrative example

Step 2 — Generate a draft, and judge it yourself

The writer gives Quillecho a topic: "a piece about why I'm cutting back on social media posting." No automated "voice-match score" is shown — just the archive next to the draft, for the writer to judge.

From the archive (excerpt)

"I used to think posting more was the job. Turns out the job was writing something worth reading, and posting was just the delivery truck. I noticed this the week I missed three days on purpose and nothing broke — the newsletter still landed, the inbox still filled. So here's the question I keep circling back to: what am I actually optimizing for?"

Draft, from Quillecho

Drafting in your voice… simulated

"I used to think showing up online every day was the job.

It isn't — the job is writing something worth someone's five minutes, and posting is just the delivery truck.

I found this out by accident, cutting back for a week and watching nothing break.

Which leaves an obvious question I can't fully answer yet: what was I actually optimizing for?"

Both blocks above are invented for this demo — no real writer's archive or output. The point being illustrated: matched rhythm and phrasing, not a copy-paste.

Illustrative example

Step 3 — A plain-language revision loop

The writer doesn't like the ending and asks for a punchier close, in their own words.

make it punchier — the ending's too soft, end on the question harder
Got it. Tightened the close — cut the hedge, kept your rhetorical-question habit:
New ending: "...watching nothing break. So: what was I actually optimizing for? Not an answer I have yet. Just the question I should've asked a year ago."
better. keep that.

Simulated exchange — illustrative example only. The revision loop targets the writer's own stated notes, not a generic "make it better" rewrite.