Stop building features that don't work.That RICE score. Was it right?From idea to impact, in one line.Linear is how you build. Preship is what you build.AI never asks “Do we really need this?”

No. Given your assumptions, it calculates the chance and range of reaching a goal like "$36,000 a month in 12 weeks." Every assumption is visible, so you can change any of them. After you deploy, it compares the prediction with real numbers and corrects the assumptions that missed.

The calculator is free to use right now. The record features aren't ready yet. We'll email you when they are.

No. Issue tracking stays in Linear. Preship handles the decision about what to build.

RICE is a score for setting the order. Preship predicts with a probability and a range, then compares that with real numbers after you ship. Whether you were right stays on the record.

Product management tools are good at collecting and organizing customer feedback. Preship handles predicting whether an idea will work and keeping the record of the result.

Yes. Pricing is per editor, and people who only read are free.

For a one-off calculation, yes. But the formula and assumptions shift with the prompt and model, which makes ideas hard to compare. Preship fixes the formula and assumptions, connects them to your real numbers, and only corrects them based on measured results. It also keeps a record of predictions and results, so comparisons get easier the longer you use it.

With numbers you type in by hand, yes. That's why Preship reads your real data, and treats the probability not as proof that something will sell, but as a way to see which assumptions drive the result.

Not necessarily. Ads, seasonality, and other changes around the same time can move revenue too. With few users, you often can't tell. Preship shows the overlap and the date a verdict becomes possible. Measuring it properly takes a randomized experiment.

Yes. For Google Sheets, just share the sheet with view access. BigQuery is read with read-only access. Query costs are billed to your Google Cloud account, so Preship estimates the data scanned before each query and won't run anything over the cap you set.

AI estimates it, and you can edit it. There's no guarantee it's right, and the range stays wide while there are few past results. We haven't been able to measure prediction accuracy yet. As measured results build up, the ranges fit your business better.

We don't measure it with the same yardstick as growth ideas, because the shorter the time frame, the worse defensive work looks. Legal compliance, payments, and protecting secrets aren't ranked; they get done first. For everything else, the odds and cost of an incident go into the assumptions, and Preship shows at what odds the work is worth doing first.

You keep using it. Your team and AI do the building as usual, while Preship handles deciding what to build and measuring the impact. Any AI works by pasting the prompt. Developer tools that support MCP, such as Claude Code, can connect directly.

Neither is required, and you connect them separately. With GitHub connected, PR and deploy times are saved automatically. Without it, you log deploy times by hand. Connect Linear only if you want to turn ideas into issues.

Yes. Instead of revenue, you can set a goal on an earlier number like signups. But with few numbers, most assumptions are guesses and the results get wide. If you have no visitor or signup numbers at all, we'd suggest finding your first users first.

We don't show one while there are only a few verdicts.

Numbers with few people behind them, like signup rate, often can't be measured. With 3,000 visitors a week, confirming the gap between 0.4% and 0.5% takes about 9 months, split half and half. Preship calculates the verdict date up front and measures a number with more people behind it, like the share of new accounts that create a first invoice. If that's still not enough, it says "Too early to tell" and suggests asking users directly.

Aggregated revenue and traffic numbers (read from Stripe, Google Analytics, BigQuery, or Google Sheets), your goals, your history of ideas and verdicts, GitHub PR and deploy records, and Linear issue status if connected. Connection credentials such as your read-only Stripe key are stored encrypted, and you can delete them at any time. We never hold your AI logins, and we don't read the contents of your source code. The calculator you fill in by hand runs entirely in your browser, and your numbers never leave it.