Copy Editing
It's 4:40 on a Thursday and four drafts are stacked behind one person who is on a job site. This agent reads them first — against your voice, your standards, your rules — and hands each one back with every change marked next to the words it replaced. Numbers, dates, names, and promises go on a flag list instead of an edit list. It never sends. Whoever was going to send it still sends it.
Yours outright — no required subscription.
Every agent is built for your business — your systems, your approval chain, your way of working. We scope it on the call.
If this sounds too technical, don't worry. We take care of everything for you.
- Private VPS deployment. This agent runs on a server that belongs to you, not a shared cloud tenant. Your proposals, client emails, and unfinished drafts stay on your box.
- A technician agent alongside it. Every hire ships as a pair: the Copy Editing Agent that reads the drafts, and a VPS-technician agent that lives on the same box and keeps it healthy — patched, backed up, monitored. You're not hiring one thing, you're hiring a small team of two.
- A style sheet built from your own writing. Setup starts with the documents you're proud of, not a template. The rules come out of those, and we name which voices are being corrected and which are being protected — because the blunt estimator who closes work should not get smoothed into everyone else.
- Wired into where drafts already live. Connected to the mail, docs, and chat your team actually writes in, so a draft doesn't have to be exported, forwarded, or pasted anywhere to get read.
- Security hardening. Private VPN, firewall, and encryption configured on your server before it reads a single client document.
- 14 days of priority support after launch, to tune it against the real mix of writers and documents your week produces.
- You own everything — the server, the agent, the code it runs on, and the style sheet itself as a file you can open.
Not sure it fits? Check fit in 90 seconds in the free assessment chat.
Who this is for
- You're a small agency where the writing is the deliverable — client updates, recaps, case studies — and quality control is an account director skimming a paste between calls. On a busy week the skim doesn't happen and the client is the first reader.
- You run a trade or contracting business where estimators build proposals by duplicating last month's file, and the previous client's name or a scope paragraph that no longer applies survives into the new one more often than anyone wants to say out loud.
- You are the last set of eyes on everything client-facing and you know it. Drafts land in your inbox after five with "can you look at this?", so you're either the bottleneck or the reason something went out unchecked.
- You paid someone to build a voice guide two years ago. It exists. Three people have opened it. The rules in it are real, and there is no moment in anybody's day where they actually get applied.
How it earns trust
The failure mode of AI in editing isn't a clumsy sentence. It's a wrong one that reads well — a claim that stops looking like something to check because the prose got fluent. This agent is built on the opposite bet: nothing it touches changes quietly, and everything it wasn't sure about is listed where you'll see it.
A diff, not a replacement
Every pass comes back as the original and the edit, change by change, so you can reject one suggestion without losing the rest. The pre-edit version is never destroyed. Each document keeps a stamped record: what came in, what went out, which suggestions a human accepted and which they rejected.
Every change carries its rule
Each edit shows the rule that triggered it, and each rule points back to where it came from — a line in your style sheet, or a real document of yours from setup. "Why did it change that?" has an answer you can read instead of a shrug about what the model felt.
Edits and flags in separate lists
Numbers, dates, commitments, unfamiliar terms of art, and anything in regulated language land in the flag list, not the edit list. It's short enough to read in twenty seconds, and it's where you'd catch the agent saying something confident and wrong.
How much of your outbound writing this covers in the first month depends on where your drafts actually live, and we'll size that with you on the call rather than promise a number up front.
Pairs well with
These share a workflow with this role. Tick any to add them to your setup.
I don't want it rewritten. I want it fixed.
That's the design. It returns your draft with changes marked against what they replaced, not a clean rewritten document you're asked to trust. You can reject one suggestion without throwing out the pass, and the pre-edit version is never destroyed. It is also tuned to leave a clean draft alone — an editor who returns fourteen suggestions on a paragraph that was fine trains everyone to stop reading suggestions.
My newest hire writes like a press release and my best estimator writes like a text message.
Both stay recognizable. It corrects against the standards you set in setup, not toward an average, and we name the voices it is not allowed to smooth. That matters more than it sounds: the short, blunt, unpunctuated writer is often the one closing work, and a general editing instinct would quietly delete the thing that was working.
A proposal went out with the last client's name still in it. Will this catch that?
It will flag it. It will not fix it. Names, numbers, dates, prices, and anything contractual, warranty, or medical go on a flag list for a human rather than an edit list. It can tell you the figure in paragraph three doesn't match the one in the table; it will not decide which one is right. And it cannot tell you whether a sentence is true — it can make "framing is complete" read well without any idea whether the framing is complete, which is exactly why those claims get flagged instead of polished.
I can't be reading every client email at nine o'clock at night. Does this replace my read?
No, and it shouldn't. It never sends, posts, or publishes anything, and it never replaces the person who signs off today — if you read every proposal before it goes out, you still read every proposal. What changes is what you're reading: a short flag list and a marked-up draft instead of a raw one. It adds a step ahead of yours; it does not remove yours.
We have a style guide. Nobody's opened it since the day we made it.
A guide that says the voice is "friendly but professional" has never settled an argument, and it never will. Setup turns your rules into a style sheet that gets applied at the moment of writing rather than read once. When someone tells the agent to stop doing something, that correction is written into the sheet with a date and a name, so the rule stops evaporating and coming back next month from a different person.
What's covered under the $1,997 setup, and what isn't?
One company, one system of record, one approval chain — a flat one-time setup, typically live about seven days from kickoff. If you have genuinely separate brands with different voices, or a second business that needs its own style sheet, that's a scoped conversation rather than a price adjustment. We'll size it on the free 15-minute call.
Hiring more than one? A department on tap — the subscription puts a build team behind every request, agent after agent.