Tell your AI tool to fix
all open client feedback.
Pensl makes all your open feedback readable by AI coding tools. Claude Code, Cursor, or any tool that supports MCP can see the full list, read every detail, and mark items done — without you copying and pasting anything.
What it looks like
One prompt. Your AI works through the queue.
you > Fix all open Pensl feedback for the agency-site project
Fetching open feedback from Pensl…
Found 3 open items.
1. Text change — Hero CTA “Get in touch” → “Book a free call”
File: components/Hero.tsx — fixing…
Done. PR opened: #41
2. Element change — Pricing card button colour
File: components/PricingCard.tsx — fixing…
Done. PR opened: #42
3. General comment — “footer feels too cramped”
No structured selector — marked for human review.
All done. 2 PRs opened, 1 left for review.
No copy-pasting. Your AI reads the feedback directly from Pensl.
What the AI sees
Every feedback item includes everything your AI needs
When your AI tool reads a feedback item from Pensl, it gets structured data — not a screenshot to interpret or a vague description to guess from.
Feedback type
Whether the client used text highlight, element click, or general pin. Tells the AI what kind of fix is possible.
Page URL
The exact page where the feedback was left. The AI knows which part of your site needs changing.
What was clicked
The CSS selector and full HTML of the element the client interacted with. Precise enough for the AI to find the matching code.
What the client said
The client's comment or the exact text replacement they typed. Plain language the AI can act on directly.
Before and after strings
For text changes: the exact original text and the replacement. No interpretation needed — just find and replace.
Screenshot
A full-page screenshot taken at the moment of feedback. Available for visual context when needed.
{
"type": "text",
"page": "https://your-client-site.com/",
"selector": "section.hero button.cta",
"before": "Get in touch",
"after": "Book a free call",
"clientComment": "Change the button text to something friendlier",
"fileHint": "components/Hero.tsx",
"screenshotUrl": "https://..."
}Setup
Three steps. No ongoing maintenance.
Sign up and get your MCP endpoint URL
After signing up, your organisation gets a unique MCP endpoint. It's a single URL you copy once. No config files, no environment variables to manage on your end.
Add it to Claude Code, Cursor, or your MCP-compatible tool
Paste the endpoint URL into your AI tool's MCP settings. Claude Code and Cursor both support this in their configuration. Takes about two minutes.
Compatible with: Claude Code, Cursor, and any tool that supports Model Context Protocol (MCP).
Tell your AI tool to fix the feedback
You can now tell your AI tool "go fix all open Pensl feedback" or "what feedback is open for the homepage project?" and it will read directly from Pensl. No copy-pasting, no screenshots, no context switching.
Capabilities
What your AI tool can do with Pensl feedback
Four actions. That's all it takes to go from feedback queue to resolved.
List open feedback
See all open feedback items across your projects, filtered by type or status. Your AI gets the full queue at once — no pagination to deal with.
"What feedback is open for the agency-site project?"
Read a feedback item
Get the full detail of a single item: the page URL, what was clicked, what the client said, the before/after strings, and the screenshot.
"Tell me exactly what the client wants changed on feedback #14."
Mark an item as resolved
Once your AI has applied a fix, it marks the item as resolved in Pensl. Your Kanban board updates automatically. No manual status changes.
"Mark feedback #14 as resolved."
Add a comment
Your AI can leave notes on a feedback item — what it did, what it was unsure about, or questions for your team. Visible in the Pensl dashboard.
"Add a note to #14: changed button text in Hero.tsx."
Why it matters
No more copy-pasting feedback into your AI tool
Right now, getting client feedback into Claude Code or Cursor means opening the email, reading the screenshot, typing out the description, and hoping your AI understands what the client meant. With Pensl, that whole step disappears.
Without Pensl MCP
- Client sends a screenshot in email
- You describe it manually to your AI tool
- AI guesses which file to change
- You verify it got the right element
- You copy the fix back to your editor
With Pensl MCP
- Client leaves feedback on the site
- You tell your AI tool: "fix open Pensl feedback"
- AI reads the exact element and change
- AI applies the fix and opens a PR
- You review and merge
Let your AI tool work through the feedback queue while you sleep.
Connect Pensl to Claude Code or Cursor in minutes. Wake up to resolved feedback and open PRs.
Start free — get your MCP endpoint