Per-seat pricing punishes the wrong thing
Feedback tools charge by the seat, which means the price climbs exactly as your team starts caring about customers more. AI replies get metered per resolution. And your users’ voices live in someone else’s cloud, on someone else’s terms.
Harkly exists because the math was backwards. The people who should be reading feedback are the whole team — and pricing shouldn’t tax that.
Four beats, then it repeats
The product is a loop, not a board. They ask — through a board, a widget, the messenger, a survey, an email. You decide — votes weighted by the revenue behind them, a roadmap that shows only what you place on it. You build — posts hand off to Linear, GitHub or Jira and track back.
Then the beat most tools skip: they hear back. Publishing an update notifies every voter once, flips their post to Completed, and clears the badge. The request ends where it started.
Trust is a feature, not a settings page
Vote integrity, a moderation queue, PII scrubbing, a full audit log and per-user export are in the foundation — because a feedback platform that can be gamed or leaks is worse than none.
The AI plays by the same rule: answers are grounded in your help center or they don’t go out, and you bring your own key — OpenAI, Anthropic or a gateway — at cost.
Our cloud or yours
Harkly runs at harkly.app, and the same build runs in one container on your own server for teams that need the data to stay home. One-click import from Canny, Featurebase and Intercom means leaving is as easy as joining.
We’re dogfooding it across our own products before a single outside customer. The beta list is open at harkly.app.