Tell your agent exactly what to fix.
Feedbackable turns a click on your UI into a structured item. Your agent reads the page, the element, a snippet of its markup and a screenshot over a REST API.
Early access. Reviewers need no account.
A web page with two numbered pins. A comment panel beside it shows a comment from a client and an open status.
An agent cannot see what you point at.
You tell the agent "the pricing card looks wrong". It has to guess which card, which page and what "wrong" means. You paste a screenshot and hope.
No exact target.
A sentence does not name an element.
No context.
The agent does not know the page, the screen size or the console errors.
No structure.
Pasted text has no status, no priority and no owner.
One click. One structured item.
Every piece of feedback becomes an item with the fields below.
- Page URL
- The page the reviewer was on.
- Element anchors
- Several ways to find the element again.
- Markup snippet
- A piece of the page markup around the element.
- Screen and browser
- Viewport size and browser.
- Console and network
- Errors and requests captured at that moment.
- Screenshot
- What the reviewer saw.
- Status, priority, assignee, tags
- So an agent can pick the next item.
Your agent reads. You stay in charge.
Step 1: A reviewer clicks and comments.
They use the live page. They need no account.
Step 2: Feedbackable stores a structured item.
It is in the dashboard, with a status and priority.
Step 3: Your agent reads it over the REST API.
The public API gives developers and agents the items in a form they can parse.
Several anchors for one comment.
One selector breaks when a page changes. Each item stores more than one anchor, so your agent has more than one way to find the element. If all anchors fail, the screenshot and position still show where the reviewer pointed.
Built for agents and people.
Readable by agents
The public REST API returns items in a structured form.
Exact targets
A reviewer clicks the element. Nobody describes it.
Context included
URL, markup snippet, viewport, browser, console and network information.
Survives page changes
Several anchors per item.
Priority and status
Agents and people can work from the same list.
One script tag
Works on localhost and on staging behind a login.
Questions from builders.
How does my agent get the feedback?
Over the public REST API. It reads items in a structured form.
What does the agent see about the element?
Anchors for the element, a snippet of the page markup, the page URL, the screen size, the browser, console and network information, and a screenshot if the reviewer added one.
Where do I manage status and priority?
In the dashboard. You set the status, priority, assignee and tags there, and agents and people work from the same list.
Do reviewers need an account?
No. A reviewer can comment without an account and can add a name if they want to.
This button is too small to see on a phone.
open