You vibe coded your product. Now what?

Let reviewmy check your .

Five lenses over the thing you shipped, free, with the top findings from each and a count of what is held back. Paid adds legal, security and a code review of the repository behind it. Every finding quotes or measures something, then the fixes go back to the agent that built it.

Any public page works. Run a free audit of one page first if you would rather not make an account: no signup, no card.

Reviewed pages are sandboxed and served from a separate origin.

Integrated with your favorite tools

  • Claude
  • Codex
  • Cursor
  • Lovable
alderpine.com/pricing
Review complete72/100
Pricing that scales

Simple, transparent, built for teams

Starter$0Choose
Team$29/moChoose
Scale$99/moChoose
12301Scales to what? No unit03Tap target 31×28px · needs 44
4Dana Whitfield2h ago

This tier should say “per seat”. Also the button feels cramped on mobile.

Inside a review

What the seven lenses actually check.

Each lens has to quote or measure something on the page. A lens that finds nothing says nothing, no filler, no “consider adding a testimonial section”, no findings that would be true of any website.

01

Copy

Claims without units, contradictions between pages, headings that describe a category rather than a job.

02

SEO

Duplicate descriptions, missing structured data, headings out of order, canonical mistakes.

03

UX

Tap targets under 44px, invisible focus rings, forms that lose state, contrast failures with the measured ratio.

04

Speed

Real LCP and CLS on a throttled connection, with the byte weight of whatever caused it.

05

Legal

Collection without notice, missing policy links, cookie behaviour that fires before consent.

06

Security

Mixed content, exposed keys redacted to a prefix, headers you are missing, judged against OWASP.

07

Code

With the repository connected, every finding lands on a file and a line, linked straight to GitHub.

+

Senior pass

One reviewer reads every lens report and writes the do-this-next list, ranked by what it costs to leave alone.

Stakeholder review

Your stakeholders mark up the live page, not a PDF.

Send one link. They comment where the problem is, rewrite the line they don’t like, and ask for what’s missing, on the real page, with no account and nothing to install.

review-my-site.com/r/sm-7c41 → star-mountain-site.vercel.appClient reviewer is here
Star Mountain home page heroUtility territories sectionWhich rules apply sectionFunding and services section
Client reviewer
101No deadline or program name above the fold202Eyebrow could carry Energize Denver · Colorado BPS
ThreadLive · 4 open
Comment · hero
1Client reviewer2h ago

“We calm the storm” is the right tone, but nothing above the fold says which rules or when. Owners are searching for the deadline.

COLORADO COMMERCIAL BUILDINGSENERGIZE DENVER · COLORADO BPS · BENCHMARKING DUE JUNE 1
New section request · territories
4Client reviewer1h ago

The utility chips look like logos but nothing happens when you click. Add a rebate band under them naming the four towns Efficiency Works covers.

Comment · which rules apply
5Client reviewer48m ago

The Denver card hides June 1, 2028 and 2032 inside a paragraph. Those three dates should be a list an owner can scan.

6Client reviewernow

Writing about the funding box

From review to production

Connect once. The fixes go back to the tool that built the site.

Stakeholder comments and every measured finding leave as one brief. Your agent picks it up in the repository it already has open, and what comes back is a summary you can read and a branch you can merge.

Step 01Connect MCP
Step 02Prompt the fixes
Step 03Summary and version control

One server line, once per project. Claude Code, Codex, Cursor and Lovable all speak MCP, and the connection stays read-only until you allow writes.

# one line, once per project

$ claude mcp add reviewmy \

  --url https://mcp.review-my-site.com/sse \

  --project star-mountain

connected · 6 tools · read-only until you allow writes

Claude Codeready
Codexready
Cursorready
Lovableready

Ask for the brief and the agent works through it. Every edit is tied to the finding or the stakeholder comment that asked for it.

>pull the reviewmy brief for star-mountain-site.vercel.app and apply everything mechanical
Hero.tsx:18Eyebrow names both programs and the deadline+2 −1
RulesCards.tsx:44June 1, 2028 and 2032 pulled out as a list+21 −6
Territories.tsx:12Utility chips linked to their programs+9 −9
RebateBand.tsxNew section built from the request+64

While it works

Findings in the brief16
Stakeholder comments4
Applied so far12
Held for a decision3

Every edit carries the finding id that asked for it.

We write up what changed in plain words, open the branch with one commit per finding, and re-run the review against the preview build.

reviewmy/brief-sm-7c41 → mainReady to merge

Apply review brief for star-mountain-site

01Hero eyebrow names both programs and the June 1 date.
02Denver card dates lifted into a scannable list.
03Utility chips now link to each program.
04Rebate band added under the territories.
7294Re-run on the preview build

What we handle

Branch openedbrief-sm-7c41
Commits, one per finding12
Stakeholder revisions merged4 of 4
Left for you3 decisions
Previewpassing

PR #218, the stakeholder sees this same summary in plain words.

Pricing

All four review products on every plan.

What changes between tiers is how much of the metered AI work you can run, how many repositories you can connect, and whose name is on the report. The Agent Console is not one of them: it is metered and billed on its own, and no review plan includes it. Both are priced here.

Vibe Coder

$19 /month

You shipped it with an agent. Check it before the client does.

  • 40 page reviews a month
  • 8 code reviews · 1 repository
  • Unmetered share links, pins and triage
  • SEO and AEO kits
  • 150 copy rewrites · image packs bought separately

Cracked Coder

$49 /month

For the studio shipping client sites every week.

  • 150 page reviews a month
  • 30 code reviews · 5 repositories
  • Reports under your own name and colour
  • MCP access tokens for Claude Code
  • Signed webhooks, with Zapier, Make and n8n

Enterprise

$199 /month

Volume, and a person to talk to when something is wrong.

  • 600 page reviews a month
  • 120 code reviews · 25 repositories
  • 1,600 copy rewrites · image packs bought separately
  • Priority support
  • Onboarding help for your first reviews

Everything collaborative stays unmetered. Share links, pins, threads, triage, visual edits and reports never count against an allowance. Full allowances, add-ons and what happens when you run out →

Product tour

Five products. Four share one loop.

The four review products share a project, a share link and a change brief, and every plan includes all four, with the tier deciding only how much metered work you can run. The Agent Console is the fifth: its own catalogue, its own checkout, and not part of any review plan.

Seven lenses, and none of them may guess.

Copy, SEO, UX, speed, legal, security and the code itself. Each lens has to quote or measure something on the page.

  • Severity-ranked, most expensive to ignore first
  • With the repo connected, every finding lands on a file and a line
  • A senior pass writes the do-this-next list
See a review →
review-my-site.com/projects/star-mountain
Review complete72/100
Copy4 findings
SEO2 findings
UXClean
Speed1 finding
CodeHero.tsx:24 +3

They click the thing they mean.

Send one link to whoever signs off. No account, no password, no “third card from the left, below the fold”.

  • Browser, OS, viewport and a screenshot attach themselves
  • Pins survive a redeploy, or say plainly that they did not
  • Every pin is already a task you can triage
See the pins →
review-my-site.com/r/sm-7c41
Stakeholder pins on the live page1201Dates buried in the paragraph

See what your agents are actually doing.

Which files it touched, which tools it called, and where a run went sideways. Auditable after the fact, not only watchable live.

  • A timeline per run, tool calls and writes in order
  • Every diff attributed to the agent that made it
  • Flag a run and send it back as a review
Watch a live run →
review-my-site.com/agents/run-4f1c
Planner4 steps scoped
CoderWriting CouponField
ReviewerQueued

Audit log

Coderedit Cart.tsx · +18 −41.9s
Coderflagged run_tests · retried 2×22.6s
Plannerplan_created · 4 steps0.9s

Team photos, without the photographer.

Your client’s team uploads a few pictures and gets a consistent set of studio headshots for the site you just built.

  • 20, 50 or 100 finished shots per order
  • Ordered through your share link, billed to them
  • Switched on per project, off by default
See the flow →
review-my-site.com/r/sm-7c41/headshots
50 shots · 6 people · ready in 20mOrder

Found by search, quoted by assistants.

Metadata, structured data and answer-shaped content written from the page you already have, not from a template.

  • Titles, descriptions and structured data, per page
  • Answer blocks written for retrieval, not for a keyword
  • Applied as a change brief, like any other finding
See a kit →
review-my-site.com/projects/star-mountain/seo
TitleEnergize Denver Compliance & ASHRAE Audits | Star Mountain
SchemaProfessionalService · FAQPage · areaServed
MissingCanonical on /rebate, alt text on 6 logos

When is Denver benchmarking due?

June 1 each year for buildings inside Denver city limits, with targets in 2028 and 2032.

MCP for Claude Code

You get a branch, not a to-do list.

Findings and stakeholder pins become one change brief. Claude Code reads it over MCP, applies the edits in your repository and marks each one done, or you copy the brief and paste it into whatever you use.

1

Paste the domain

Rendered through our proxy on a separate origin. Nothing to install on the site, nothing for you to deploy first.

https://alderpine.com/pricing
2

Connect the repository

Optional. A URL alone gives a full review of the page; the repo adds a file and a line to every finding.

alderpine-co/marketing-site
3

Hand it to the agent

One MCP token for the account. Claude reads the open requests, revises, and closes them out from its side.

claude mcp add reviewmy

Coverage

Whatever built it, we review it.

The review runs against the rendered page, so the stack does not matter. The repository half is where the toolchain does, these are the ones we read natively.

  • Claude
  • Codex
  • Cursor
  • Lovable
  • GitHub
  • Supabase

Coding agents

Read the outstanding requests and mark them done over MCP.

  • Claude Code
  • Codex
  • Cursor
  • GitHub Copilot
  • Windsurf

App builders

Export or connect the repo and the code pass comes with it.

  • Lovable
  • v0
  • Bolt
  • Replit
  • Base44

Repos and hosts

Where the file-and-line links point, and where deploys are noticed.

  • GitHub
  • GitLab
  • Vercel
  • Netlify
  • Cloudflare Pages

Stacks and platforms

Reviewed from the rendered page, so no plugin and no build hook.

  • Next.js
  • Astro
  • WordPress
  • Webflow
  • Shopify

Marks belong to their owners and identify the tools reviewmy reads. Send any others you have permission to use and they go in the strip.

Measured, not claimed

What a review actually returns.

7Model calls in a full review
4Of them on the free audit, no card
0Accounts your client has to create

Point it at the last thing you shipped.

One page, no account, no card. If it finds nothing, it will tell you that too.