AI ships faster than your PRD can keep up.
You're the first person called when the product doesn't match what the customer wanted. Spec Kitty turns your Product Requirements Document (PRD) into the artifact from which your team builds. Your customers get what they asked for, and you spend less time fielding the calls when they don't. More forward-looking discovery. Less reactive cleanup.
Heads up: Spec Kitty is a Python program that runs on the command line. You will also need an AI subscription like Claude Code, ChatGPT, Cursor, or GitHub Copilot.
Workshops are for engineering-leadership-driven team rollouts. For just you, the early access above is the right starting point.
"Artificial Intelligence (AI) ships faster than your PRD can keep up."
Your PRD isn't the bottleneck. It's what's been outpaced.
Your engineering team adopted Cursor (or Claude Code, or Gemini) somewhere between Q3 2024 and last sprint. They're 2–5× faster on implementation. You write the same number of PRDs at the same speed you always did. Now your PRD is what's outpaced.
That's not a bottleneck — that's a missing layer between what you wrote and what the agent built. Spec Kitty makes the spec — your PRD — the artifact agents read before they generate code. Your PRD doesn't sit in Notion or Confluence for the engineer to interpret; it's the prompt the engineer's agent works from.
"By the time you see it, three sprints of work are wrong."
Silent spec drift — the gap between agent shipped and Product Manager noticed it shipped wrong.
Tests pass. Code looks fine. The standup says everything is on track. Then Quality Assurance (QA), or worse, a customer call: the wheels came off and you were the last to know. Engineers have a name for it (drift detection). Product Managers usually just call it the pizza you didn't order.
Spec Kitty turns the gap into an audit trail. Every change to the spec is timestamped. Every implementation is reviewed against it before it lands. You see drift the day it happens, not three sprints later in a 1:1 with your Chief Product Officer.
"What's left of my job that AI can't do?"
The human parts. Spec, intent, decision trail. Make those auditable.
You've read the headlines. Your CEO has read them. The Slack thread that started at 11pm and didn't end until 2am — you were in it. Whatever you call it — the existential question, the seat-at-the-table question — it's the weather your team is shipping in.
Spec Kitty's answer is structural, not motivational. The brief, the bet, the decision the Product Manager made — that shipped to the customer. Those are the human parts agents are still bad at. Make those the auditable artifact and the seat at the table re-justifies itself.
We're not going to tell you AI won't replace Product Managers. We're going to give you the tool that makes the parts it can't do visible to the people deciding.
Frequently asked questions
Is this just Productboard or ChatPRD with a different sticker?
No. Productboard is roadmapping and feedback intake. ChatPRD writes PRDs. Spec Kitty is the layer that makes your PRD the artifact agents (and engineers) read before code is generated — and tracks the drift between what was specced and what shipped. If you already use ChatPRD to draft, paste its output into Spec Kitty as your starting spec. They're complements, not replacements.
Should I send this to my CTO before getting started?
Not required. The Product Manager workflow is built for you to start today, on your own PRDs, without anyone's sign-off. If your CTO is already asking about spec-driven development, forwarding this page to them is also a fine path. Most teams begin with one Product Manager and bring engineering in once the workflow proves itself.
What if my engineering team isn't using AI coding tools yet?
Then this isn't for you yet — honestly. Spec Kitty solves the speed-differential pain that opens up after your engineers adopt AI coding tools. If they haven't, your PRD-to-ship workflow is probably fine for now. Bookmark us for the day the first Cursor pull request lands without a ticket.
What happens to the PRDs I paste in?
Your PRDs stay in your account. We don't train AI models on them. We don't share them with third parties. We use them only to give you the structured spec and drift-detection output — the product. Delete your account and we delete your content within 30 days. Full detail in our privacy policy.
Built by people who shipped at GitHub and Red Hat
Creator of Spec Kitty. Long open-source background; ex-Acquia / Platform.sh; building the layer between product planning and agentic delivery.
VP of Engineering Operations and Chief of Staff to the CTO at GitHub. Chief Agilist at Red Hat ($2B Red Hat Enterprise Linux business). 30 years in tech, zero patience for unnecessary meetings.
Stop being the last to know your spec drifted.
Join the waitlistWe'll reach out when it's ready.