Why we started cyql
Software teams don't have a code-writing problem. They have a follow-through problem. We built cyql to close the gap.
By The cyql team
Every engineering team we talked to had the same list. The dependency bump nobody gets to. The flaky test that sits open for three months. The migration that has been in the backlog for a quarter. The linting rule someone proposed in a PR review six months ago that never got enforced.
None of this work is hard. It is well-defined, it is important, and it never gets prioritized because there is always something more urgent. The backlog grows. Technical debt compounds. The things you know you should do stay undone.
AI coding assistants make individual engineers faster. But they do not close the gap. They help you write code when you sit down to write it — they do not sit down for you. The follow-through problem is not about typing speed.
cyql is built for the follow-through. You describe a task in plain language. cyql plans it, dispatches a fleet of isolated agents, writes and reviews the code, and opens a pull request. The loop runs until the work is done or it knows it needs to ask you something. You stay in control of the decisions that matter; the agents handle the work that should not need your attention.
The name is a nod to how it works. cyql ≈ cycle. Every task runs the same loop: plan, dispatch, code, review, ship. And when something fails, the loop replans instead of giving up. We built it this way because follow-through is not a one-shot problem — it is a loop that runs until done.