marca
Start free trial →
marca.dev · the input layer for ai coding

The red pen for AI coding.

Your agent writes code fast. The bottleneck moved to the spec. Click any issue on a live page — Marca captures the screenshot, selector, URL, and viewport, then exports a fix brief you paste straight into Cursor or Claude Code.

Start 14-day trialSee pricing

reviewers always free · canvas on every plan · from $39/mo

staging.acme.dev/pricing

Simple, honest pricing.

Pick the tier that fits your team. No hidden fees.

Free
$0/mo
Get started
Pro
$29/mo
Get started
Team
$99/mo
Get started
jess · designer
Pro CTA contrast fails AA — 2.9:1 on this background.
3 pins · 2 reviewers · live

Stop translating Looms, Figma threads, and Slack screenshots into tickets.
Mark the page. Ship the fix.

From someone noticed to the agent shipped — in three steps.

One script tag on staging. Anyone you invite can mark — nothing to install, free one-click sign-in for reviewers.

01 / drop in

One script tag on staging.

Marca attaches a floating widget to your site. No build step. No SDK. Strip it before prod, or leave it for QA.

<script src="https://marca.dev/m.js"
data-project="YOUR_PROJECT"></script>
02 / mark anything

Click. Comment. Done.

Reviewers click any element and leave a note. Marca captures the selector, URL, viewport, and screenshot automatically.

03 / export

Paste straight into the agent.

Select what matters, hit Export. You get a markdown brief with selectors and screenshots, ranked by priority. Paste into Cursor or Claude Code.

# Marca brief
## 1. CTA contrast fails
## 2. Logo links to /home
The export

The brief your agent reads.

Every pin exports as clean markdown — the selector, the viewport, the screenshot, and what good looks like. No screen-recording, no re-explaining. Paste it into Cursor or Claude Code and the agent has everything it needs to ship.

marca-brief.md
## Fix: Pro CTA contrast fails WCAG AA
Page: /pricing
Element: button[data-plan="pro"]
Viewport: 1440 × 900
Issue: Text contrast is 2.9:1 on the current background.
Expected: Adjust button / text colors to meet 4.5:1.
Priority: High
Screenshot: attached
# reviewer: jess · designer · 10:02

Before Marca. After Marca.

Same bug. Same team. Different loop.

Before

Translate, ticket, ship.

  1. monQA records Loom of the bug
  2. tuePM watches Loom, writes ticket
  3. wedEngineer reads ticket, asks for details
  4. thuEngineer reproduces, writes spec for agent
  5. friAgent ships fix
cycle time5 days
After

Mark, export, ship.

  1. 10:02QA clicks the bug on staging
  2. 10:03Marca captures selector + screenshot
  3. 10:05Engineer exports markdown brief
  4. 10:06Pasted into Claude Code
  5. 10:14Agent ships fix
cycle time12 minutes
Who it’s for

Built for teams shipping with AI.

Anyone who reviews a live staging app and turns what they find into work for Cursor, Claude Code, Linear, or GitHub.

Safe for staging

You decide where it runs.

Staging-first

Built for review and staging environments. Strip the one script tag before production, or leave it on for ongoing QA.

You control access

You decide which sites run the widget, who can mark pages, and what gets exported. Reviewers only see what you invite them to.

No source access

Marca sees the rendered page your reviewers see — element, URL, viewport, screenshot. It never connects to your repo or reads source code.

Pay for the work. Not the people.

Marca prices by pins — the unit of work — not by seats. Engineers don't pay for their PM clicking a button. Reviewers don't pay at all.

Start 14-day trial →Compare plans

Drop it on staging before your next sprint.

The agent gets sharper. The team moves faster. The fixes ship.

Start 14-day trial →Talk to us

card required · cancel in one click · reviewers always free

Swap YOUR_PROJECT for the token in your project’s Embed tab — you’ll get a ready-to-paste tag after signup.

Marca — the red pen for AI coding