Catch every bug.
Ship to green.

Val is a QA agent that lives inside your coding assistant. Point it at your app, it clicks through every flow, finds what's broken, and loops with your AI until everything's green.

100% MCPbring your own key$4.99/mo
Get ValSee what it catches
Built for
  • Vibecoders
  • Indie hackers
  • Solo founders
  • Cursor
  • Lovable
  • Bolt
  • Replit
  • v0
  • Windsurf
  • Claude Code
  • Side projects
  • Hackathons
What it catches

Every bug
a user would hit.

Broken links & 404s

Every link followed, every route loaded. Val flags the ones that dead-end on a 404 or a blank page before a visitor ever hits them.

Console errors

Uncaught exceptions, failed fetches, red console spew. Val watches the console on every page and reports what the browser is screaming about.

Dead buttons & forms

The button that does nothing, the form that never submits, the modal that will not close. Val clicks everything and tells you what did not respond.

Hidden & clipped text

Text trapped behind a container, cut off at the edge, or invisible on its own background. Val reads what a human cannot and surfaces it.

Broken layouts

Overflowing containers, elements stacked wrong, a mobile view that falls apart. Val checks how it actually renders across sizes.

Broken images & assets

The missing logo, the 404'd screenshot, the icon that never loaded. Val catches every asset that failed to show up.

Dead-end flows

Sign up, then nowhere. Checkout that loops. Val walks your key journeys end to end and finds exactly where they break.

Janky first paint

Pages that flash, shift, or hang before they settle. Val notices the rough first impression your users feel but cannot name.

Regressions

Fixed one thing, broke another? After every fix Val re-runs the whole sweep so a patch never quietly takes something else down.

The loop

Detect. Fix.
Verify. Repeat.

01

Detect

Val drives a real browser through your app, on localhost or live, and clicks every link, flow and screen to find what's actually broken.

02

Fix

Each finding goes straight to your coding agent as a precise, actionable instruction. It edits the code in your own editor, on your own key.

03

Verify

Val re-runs the exact check, then the whole critical path, to confirm the fix held and nothing else broke. If it did, the loop runs again.

Detect, fix, verify, repeat, until everything is green. You just say yes.

Lives in your
coding agent.

Val is 100% MCP. Drop it into the coding assistant you already use, point it at your app, and it becomes a QA agent your AI can drive. No new dashboard, no new login.

Claude CodeCursorWindsurfClaude Desktop

Bring your own Anthropic or OpenAI key. Val runs on your tokens, never ours, so a deep run costs you cents and stays private.

Request early access
mcp.json
{
  "mcpServers": {
    "val": {
      "command": "npx",
      "args": ["-y", "@nyx/val-mcp"]
    }
  }
}

Ship with
confidence.

Pricing

One agent.
Five bucks.

Bring your own model key. We charge for the agent and the check library, never your tokens.

Full access
$4.99 / month
  • The full check library, updated weekly
  • The self-fixing detect, fix, verify loop
  • Unlimited runs, on your own key
  • Localhost and live URLs
  • Every supported coding agent
Request early access

Anything else
on your mind?

What does “100% MCP” actually mean?+
Val ships as an MCP server, not a website. You connect it to a coding assistant that speaks MCP (Claude Code, Cursor, Windsurf, Claude Desktop) and Val becomes a set of QA tools your agent can call: crawl the app, run checks, verify a fix. There is no separate dashboard to learn.
Who pays for the AI tokens?+
You do, with your own Anthropic or OpenAI key, and that is the point. Val never spends tokens through us, so your usage stays private and a thorough run costs you a few cents. The $4.99 is for the agent and the check library, not the compute.
Can Val test my app while it is still on localhost?+
Yes. Because Val runs through your own machine and agent, it can drive localhost:3000 while you build, as well as any deployed URL. Tools that only scan a public link can never see your work in progress.
Does Val fix the bugs, or just report them?+
Both. Val hands each finding to your coding agent as a precise instruction, the agent edits the code, and Val re-runs the check to confirm it is fixed. It loops until everything is green. You approve, the agent does the work.
Is this security testing?+
No. Val is strictly about user experience: broken links, 404s, dead buttons, hidden text, busted layouts, broken flows. It does not do penetration testing or vulnerability scanning, and it never claims to. One job, done well.
Which coding agents does it work with?+
Anything that supports MCP. Today that means Claude Code, Cursor, Windsurf, and Claude Desktop, with more as the ecosystem grows. If your assistant speaks MCP, Val plugs in.
Will it break my code?+
Fixes land in your editor for you to approve, never auto-deployed. After every change Val re-runs the full sweep to catch any regression, and you can undo anything in one step. Val proposes, you decide.
What does the $4.99 get me?+
Everything. There is no free tier and no usage caps from us: the full, continuously updated check library, the self-fixing loop that re-tests until green, and unlimited runs on localhost or live. The only thing you pay on top is your own model tokens, on your own key.