Quantify client feedback.
Reviewers pin comments on your preview site. Each one arrives with the exact element, a screenshot, the viewport, and the browser.
Open source (AGPL) · Self-hosted with Docker
What happens when someone leaves a comment.
No screenshots pasted into Slack, no thread spent working out which button they meant.
- 1
A reviewer leaves a comment
They click the element, type what is wrong, and submit. The widget is already on your preview site, so there is nothing to install and no account to create.
- 2
It arrives with context
Headline too soft/pricing.hero h11440×900The comment lands in your inbox with the page URL, the CSS selector, a screenshot, the viewport, and the browser — enough to reproduce it without asking.
- 3
You triage it
Reply, set a status, group duplicates, and resolve. “Open in preview” jumps straight back to the element on the live page.
Every comment carries what you need to reproduce it.
Each comment is anchored to the element it's about and arrives with the details below. “Open in preview” takes you straight back to it on the live page.
- Page
- /pricing
- Element
- .hero h1
- XPath
- /html/body/main/section[1]/h1
- Viewport
- 1440 × 900
- Browser
- Chrome 124 · macOS
- Screenshot
- Captured
- Reviewer
- Mia (no account)
It fits how agencies already work.
You already share a preview link with clients. Put Tack's widget on it, and every stakeholder can leave feedback in place — they click, type, and you get it with full context. Nothing for them to install, no account to create. Every comment and reply stays in one inbox you control.
Self-host the whole thing.
Tack is AGPL open source. Run it on your own infrastructure with Docker and claim the instance with your email — your data stays with you. The AI Inbox is optional and uses your own API key.
$ docker compose up --buildOpen the app, claim it with your email, and add the widget to your preview site.