Comparison
Looking for a MCPFeedback alternative? Here is the honest version.
- MCPFeedback
- $9.99 per site per month, 1,000 submissions included
- UserWants
- One plan, unlimited projects and widgets, nothing metered
What they are
MCPFeedback in one sentence
Straight answer
Where each one wins
Pick MCPFeedback for
Screenshot annotation
They give the reporter a pen, arrow, rectangle and text tool to mark up the region they select. UserWants captures the viewport and offers no drawing at all.
Region select
The reporter chooses the part of the page to capture. UserWants always takes the visible viewport.
Per-site pricing suits one site
If you only ever run one site, $9.99 a month is less than a flat plan built for people running several.
Pick UserWants for
The evidence arrives without being asked for
Every bug report carries the uncaught JavaScript errors with their stack frames and the last 50 things the person did — clicks, form submits, SPA navigations — collected from page load, not from when they opened the widget.
Unlimited projects on one plan
Agencies and anyone running more than two sites pay once. There is no per-site multiplication and no submission cap to watch.
Role-aware MCP access
The editor authenticates as a person through OAuth and inherits that person’s project role, so a teammate who cannot change a status in the admin panel is not offered the tool in their editor either.
Side by side
Feature comparison
| Feature | UserWants | MCPFeedback |
|---|---|---|
| Install | One script tag | One script tag |
| Reporter needs an account | No | No |
| Screenshot | Viewport, automatic | Region select, annotated |
| Annotation tools | No | Pen, arrow, box, text |
| JavaScript errors + stack traces | Up to 10, automatic | Console capture |
| User steps before the report | Last 50, automatic | Not documented |
| MCP server | Yes, OAuth, 3 tools | Yes |
| Feature requests as well as bugs | Yes, one inbox | Yes |
| Pricing | Flat, unlimited sites | $9.99 per site / mo |
| Submission cap | None | 1,000 / month |
Verdict
Which one to pick
If what your reporters need is to circle the broken thing and write on it, pick MCPFeedback — annotation is the one place they are clearly ahead, and no amount of diagnostics replaces a person drawing an arrow. Pick UserWants when the reports come from people who cannot describe a bug well, because the stack trace and the click trail arrive whether or not they explain anything, and when you run more than one site.
UserWants is for
Teams running several sites whose reporters are users, not testers
MCPFeedback is for
A single site where reporters are briefed to mark up what they see
FAQ
Before you switch
- Can I migrate my existing reports?
- Not automatically. Both tools export, but there is no importer — in practice teams run both for a sprint and switch the script tag when the new inbox has enough in it.
- Do both work with Claude Code and Cursor?
- Yes. Both expose an HTTP MCP server over OAuth, so any editor that speaks the protocol connects to either one — Claude Code, Cursor, VS Code and Zed included. The configuration is the endpoint URL and nothing else, with no token to create or rotate.
Try it beside MCPFeedback.
Paste the tag on the same site and run both for a sprint. The reports arrive in parallel and nothing about your current setup changes.