A product manager describes the change on a live copy of the product. The AI reads a one-line request or a messy ticket, and edits the project on your computer. A developer looks once, then puts it live or leaves it there.
Can the big button on the home page say “Book a call” and open our Calendly?
Working on it…
AI
Priya
Can the big button on the home page say “Book a call”…
The main button on the home page says “Get started” and goes to the signup page. I changed it to “Book a call” and pointed it at the Calendly link you sent.
If you don't see the change in the app, refresh this page.
Alexdev
Looks right. I’ll publish it.
Message… (paste a screenshot or attach a file)
Paste a screenshot or attach PDF, Word, or Excel. Type @ to tag someone.
The change itself is quick. The messages around it are not: which page, which words, which comment.
Where a developer’s time goes
A one-line request, and a messy ticket. Both follow the same path.
“Make the button say Book a call.”
Today
The product manager writes that in Slack.
The developer asks which page. A screenshot comes back.
Then someone opens the project, finds the button, and sends a link to check.
On Butterneck
The product manager types that sentence on a live copy of the product.
The AI finds the button and changes it.
The developer opens the page and decides whether it goes live.
A ticket full of comments and files
Today
The ticket has a description, a pile of comments, a spreadsheet, and screenshots.
A developer reads all of it and figures out which comment is the real request.
Only then does anyone open the project.
On Butterneck
Add the label butterneck and assign the ticket.
Jira starts the chat. The AI reads the ticket and the files, then makes the change.
The developer reviews it and decides whether it goes live.
Why this is worth doing now
AI can already make the change
It can edit a real project, working with your team.
The slow part is the back-and-forth
Which page. Which screenshot. Which comment is the real one.
Passing that between the product manager, the AI, and the project takes longer than the change.
That is the time this saves
The change itself is usually small.
What takes the week is getting the request to the project.
How this is different
AI in the code editor
A developer still has to open the ticket and paste it in.
A ticket sitting in Jira
A developer still has to carry that ticket over to the project.
Butterneck
A product manager asks on the live product, or a labeled Jira ticket starts the chat.
The AI edits the project.
A developer publishes it.
How it works
Connect the product, let them ask, and decide what goes live.
01
Connect the product
Create a workspace.
A developer installs a small connector on a computer that already has the project. That computer needs Node.js 22 or newer, and a coding assistant already signed in: Cursor, Claude Code, or Antigravity.
They point it at the project and start the app. Everyone else then opens a live copy.
02
Hand the request to the AI
People on the product team describe the change in chat, on that live copy.
Or add the label butterneck to a Jira ticket and assign it.
That ticket starts the chat on its own.
03
You decide what goes live
Read the conversation and look at the app.
Publish it the way you already do.
If the change is wrong, leave it. Nothing goes live until you say so.
A Jira ticket can start the chat
Connect once. Label the ticket. The preview link comes back on it.
Connect Jira once
Sign in with Jira Cloud and choose your site.
Link a Jira project to your app.
Paste Butterneck’s address into Jira, so Jira can tell Butterneck when a ticket is created or updated.
A labeled ticket starts the chat
Add the label butterneck.
Assign the ticket to someone whose computer has the project.
Butterneck opens a private copy of the project, named after the ticket.
The title, description, comments, and attached files become the request.
The link goes back on the ticket
When the AI is done, a short note and a link to the preview are added to the ticket.
The full conversation stays on the live copy.
Updating that same ticket later does not open a second copy.
Access
You choose who can ask.
The link opens the preview. You decide whose message starts the AI.
You choose who can askThe link opens the preview. You decide whose message starts the AI: a list of people, or anyone with the link.
Looking does not create an accountSomeone can open the preview. They can send a message only if you already allow them.
They can ask to be let inIf the link is closed, a visitor asks for access. You allow it or decline it.
Your call
You decide what goes live.
The AI edits the project on your computer. Your team publishes it.
Your project onlyThe AI stays in the project you connected. Secret settings files (.env) stay blocked.
The conversation stays with the changeThe request, the screenshots, and the replies stay on the live copy of the app.
Your team publishesWhen the change looks right, you publish it the way you already do. Until then it stays on the computer.
Built for product teams
A team that builds a product and keeps it running.
The moment this helps: passing requests around takes more of the week than the work itself.
A product manager describes the change. A developer decides what goes live.
The project lives on a computer you control.
A company that looks after a client’s website can connect that project the same way.
Pricing
Start with your product team
Starter is free for one developer and one person who requests changes. Pro is priced per person each month. A larger team is a conversation with us.
Starter
One product, and the person who requests the changes.
