Nothing to paste? — a wishlist brief for an online plant shop: real requirements, a Figma link, two roles, one undecided price question and a legal TBD. Or see a saved backlog rendered with no run and no charge: or .
Paste the feature brief
The feature description, the requirements, the thread where it was argued out. The scanner sizes it up free in your browser: roles named, requirement-shaped lines, design links. No sign-in needed for that.
Say what the product is
Name the product and its users, and the stories are written for them — roles come from your brief, and what the brief does not decide becomes an open question, never a guess.
Take them to the board
Every story independently testable, 3–6 criteria each, sequencing notes for the sprint, and a role-by-role grounding check against your own brief. Copy, download markdown, or export the Jira CSV.
Does it invent requirements the brief does not carry?
No. Every role, behaviour, constraint and design reference in the stories has to come from the material you paste or the context you type. Limits, flows and platform constraints are never filled in from plausible guesses, personas the brief never names are not invented, and a gap in the brief is named as a gap — under Open questions, where someone can answer it. After the run the app cross-checks the story roles against your paste and flags any role the brief never carried.
What makes a story sprint-ready here?
The 3 C’s and INVEST, enforced by the output contract: each story has a plain-language title (Card), a one-sentence As a / I want / so that description whose benefit is a real outcome (Conversation), and 3–6 acceptance criteria each checkable without interpretation (Confirmation). Stories are split to be independent and sized for one sprint; when one sentence hides three stories, it is split, and when a split would create a story with nothing testable, it stays whole and the sequencing notes say so.
Do I have to sign in?
Not to size up a brief. The scan — roles named, requirement-shaped lines, design links, concrete limits, untestable wording — runs entirely in your browser, free, with nothing leaving the page. Writing the stories is a metered model run, so that step needs an account, and the reserve is shown before you press the button.
What if my brief is very long?
Material over 60,000 characters is clipped from the middle, never the end — the background at the top and the late decisions, edge cases and design links at the bottom are both sent in full. The cut is announced in the material itself, so the stories report the gap as unknown rather than guessing at what was in it, and the app tells you how much was removed.
What can I do with the stories afterwards?
Copy the whole backlog, copy any single story as a self-contained card or as a markdown checklist, download the set as markdown, or export a CSV whose Summary, Description and Acceptance Criteria columns import straight into Jira or Linear. Every story card also carries a prefilled Draft in Linear link, and naming a GitHub repo adds a prefilled New GitHub issue link per story — title, description and the acceptance criteria as a task-list checklist already filled in. The open questions have their own copy button, ready to go back to the thread they came from. Every one of those is built in your browser from a result you already ran, so none of them costs another run. Past backlogs are kept on your account, and when the brief changes you can re-run it and see which stories survived, which were dropped, and which are new.