ShipSprint vs Height

What Height was, and why this reads differently now

Height launched in 2018, built by Michaël Villar, a former Stripe engineer, and went on to raise more than $18 million, including a Series A led by Redpoint Ventures. It was a task and project tool with Kanban boards, a timeline view, chat, and one of the more polished natural-language task-capture systems around: type a sentence, get a task with the right assignee and date parsed out of it. In its final version, Height 2.0, launched in October 2024, the company pitched it as "the first autonomous project collaboration tool," meaning the software itself tried to infer and update task status from connected activity like GitHub commits and chat, rather than requiring someone to move a card by hand.

Michaël Villar announced the company was winding down in March 2025, and Height stopped operating on September 24, 2025. The public announcement thanked customers for the journey without laying out a detailed reason for the shutdown, which is fairly typical of how these things get announced. Companies close for all sorts of reasons that have nothing to do with whether the product itself was any good, and Height had real fans right up to the end.

ShipSprint takes a different structural bet on the same underlying problem: how does a team know what's actually happening on a board without someone typing it in? Instead of an AI inferring status from signals it's watching, ShipSprint uses a WIP limit on a column as a hard, mechanical stop, and a shared triage inbox as the intake gate before anything becomes a card at all. GitHub sync ties a specific, deterministic trigger, a branch tied to a card, a merged pull request, to card movement, rather than an AI reading arbitrary activity and guessing at what it means.

Neither approach is obviously "more correct." Height's bet was that AI inference could remove manual board maintenance entirely. ShipSprint's bet is that a small number of hard, predictable mechanisms beat a smarter but less predictable one. Since there's no live Height to test that bet against anymore, this is a comparison of philosophy, not a features table.

Where Height's bet was genuinely worth taking seriously

It's worth saying plainly: the idea Height was chasing, an AI that reads your team's actual activity and keeps status current without anyone touching a board, is a real and appealing goal, and it's different from anything ShipSprint attempts. Its natural-language capture and clean interface were well regarded by people who used it, and the shutdown itself doesn't mean the underlying product idea was wrong. Well-funded startups close for reasons that often have more to do with the business than the product.

If a team specifically wants that autonomous, AI-inferred-status approach today, they'd need to look at tools still actively building toward it. ShipSprint isn't one of them, deliberately. Forecasts here come from a team's own measured velocity as sprints complete, not from an AI's read of chat and commit history, and cards move because someone moved them, or because a specific, predictable GitHub event triggered it. That's a narrower promise than what Height was attempting, and it's the trade ShipSprint makes on purpose.