Huncher
Huncher scans a live website and surfaces the conversion, design and SEO leaks costing it leads, then writes a ready-to-paste prompt for each one so an AI coding agent can apply the fix.
What is Huncher?
Huncher is a web-based site audit tool built on a single premise: a diagnosis is worth only what you can do with it. You give it a URL; it crawls up to 25 pages, rebuilds them in a sandbox and inspects the DOM as it actually renders, meaning real contrast, real layout and real copy rather than an inference drawn from raw HTML. What comes back is a score out of 100, a count of leaks, and sub-scores per dimension; the example on the homepage breaks a site down into Trust, CTA, Hero and Copy, each marked out of 20.
The same scan is then read four ways. Conversion surfaces the CRO and UX problems quietly costing signups, such as weak value propositions, CTAs buried below the fold, mobile fold issues, missing lead capture and absent social proof, each ranked by ICE. SEO runs more than thirty technical checks across every page: titles, meta descriptions, canonicals, structured data, social tags, indexability, alt text, sitemaps and Core Web Vitals. Design is a product-designer pass on hierarchy, spacing, type scale and contrast, with a stated speciality of spotting the unmistakable AI-template look, from Inter on white to the purple gradient hero and the stock gradients, and proposing fixes built on your own brand tokens instead of another template. Competitors produces a feature-by-feature capability matrix against close rivals, with a build prompt for each gap worth closing.
The distinguishing output is the prompt. Every leak ships with a self-contained, production-ready prompt you paste into an AI coding agent, which then edits the real repository, with a built-in check meant to skip false positives. Leaks are also shown in place on the live page, where they occur. An Ask Huncher copilot answers questions grounded in your specific scan rather than in generic advice, and an MCP server lets Cursor, Claude or v0 fetch the leaks and apply the fixes with no copy-paste at all.
Two limits come from the vendor itself: results are informational and guarantee no particular outcome, and you may only scan sites you own or are authorised to analyse. The product is also very young, its domain having been registered on 13 July 2026, the date every legal page carries.
What it does
- Crawl up to 25 pages of a live site and rebuild them in a sandbox to inspect the actually rendered DOM
- Surface the conversion leaks costing signups and rank each one by ICE, for impact, confidence and ease
- Write a self-contained, production-ready fix prompt for every leak it finds
- Run more than thirty technical SEO checks per page, from titles and canonicals to Core Web Vitals
- Audit visual craft, covering hierarchy, spacing, type scale, contrast and the generic AI-template look
- Build a capability matrix against close competitors and propose a build prompt for each gap
- Expose the whole scan to Cursor, Claude or v0 over an MCP server so the agent applies fixes itself
When to use Huncher / When not to
A quick filter to help you decide if Huncher is the right fit.
When to use Huncher
- Founders and solo operators running a live site with no in-house CRO or design specialist
- Growth and performance marketers who need to know which page element to fix first
- Teams whose site was built with an AI generator such as Lovable, Bolt, v0 or Replit
- Product designers auditing hierarchy, contrast and typography across a whole site at once
- Developers already working in Cursor, Claude or v0 who want fixes they can apply over MCP
When not to use Huncher
- Anyone needing A/B testing or measurement: Huncher diagnoses and proposes, it does not measure what a fix actually changed
- Sites larger than the 25-page ceiling that applies to every scan
- Buyers who must know the price before creating an account, since no rate is published anywhere
- Organisations that require a signed DPA or a security certification such as SOC 2 or ISO 27001
- Anyone wanting to analyse sites they neither own nor are authorised to scan, which the terms forbid
How to use Huncher
A typical end-to-end flow, from setup to results.
- Confirm you own the site or are authorised to analyse it, which the terms make your responsibility
- Create an account on the sign-up page; the first page you scan costs nothing
- Paste the URL of the site you want analysed
- Let Huncher crawl up to 25 pages and rebuild them in its sandbox
- Read the overall score out of 100, the leak count and the sub-scores per dimension
- Work through the four views in turn: conversion, design, SEO and competitor gaps
- Inspect each leak where it sits on the live page to see the problem in context
- Copy the fix prompt attached to a leak and paste it into your AI coding agent
- Or connect that agent to Huncher's MCP server so it fetches the leaks and applies fixes directly
- Ask the Ask Huncher copilot about any leak, score or fix that needs explaining
Pros & Cons
Pros
- The output is actionable: each finding ends in an executable prompt rather than a recommendation to interpret
- Inspection happens on the rendered DOM in a sandbox, so contrast and layout reflect what visitors actually see
- Four angles, being conversion, design, SEO and competitors, from a single scan
- ICE ranking answers the question most audits leave open, namely what to fix first
- The MCP server removes the copy-paste step entirely for teams already working in Cursor, Claude or v0
- The first page is free, so the quality of the output can be judged before any commitment
- Sub-processors are named and few, and the terms leave ownership of results and prompts with the user
Cons
- No public pricing at all: no rate card, no credit-pack size, no currency, so an account is required to see any figure
- The publisher is identified by brand only, with no legal form, no postal address and terms pointing to an unnamed operating entity
- Governing law and jurisdiction are left unnamed, so it is unclear where a dispute would be heard
- No DPA is published or offered, and no security certification such as SOC 2 or ISO 27001 is claimed
- No hosting country or region is named anywhere in the privacy policy
- Scans are capped at 25 pages, which rules out large sites
- The product was barely two months old at review time, with no published customer, case study or social presence
Pricing & Plans
Huncher publishes no rate card. A free entry point exists and is stated plainly on the site: the first page you scan costs nothing. Beyond that, the terms specify that some features are free while others consume credits or require a paid plan, and that pricing and limits are shown inside the product rather than on the public site, where no pricing page exists. No amount, currency or billing unit was public as of 7 September 2026, so the lowest paid price point cannot be stated here. Fees are described as non-refundable except where the law requires otherwise, and material changes to paid terms are to be notified in advance.
Data, GDPR & hosting
A consolidated view of how Huncher handles your data.
GDPR overview
The acronym GDPR appears nowhere on the site, and Huncher never claims compliance in so many words. What it does publish, however, is written in GDPR vocabulary: named legal bases, being performance of the contract, legitimate interest and consent where required; a rights section covering access, rectification, portability, erasure, objection, restriction and withdrawal of consent; and a transfers clause invoking standard contractual clauses. A short sub-processor list is published. No Article 27 EU representative is designated and no data protection officer is named, so requests go to the contact form or to the single published address. The response commitment is worded as within a reasonable timeframe, with no figure attached. The policy carrying all of this is dated 13 July 2026.
Who owns the data?
The terms draw a clean line. Huncher owns its own software, brand and site content, while you keep ownership of your website content and of everything a scan produces for you: the results themselves and the fix prompts, which you are explicitly free to use to improve your site. On the personal-data side, the privacy policy states that Huncher does not sell, rent or trade your personal data. Crawled page content is passed to a named AI sub-processor for analysis, limited to what that analysis requires. No third party is granted rights over your material beyond what running the service demands.
Reuse rights
You may reuse everything Huncher generates for you without asking permission: the terms state that you retain ownership of the scan results and fix prompts, and that you are free to use them to improve your site. There is no licence-back clause, no attribution requirement and no restriction on commercial use of that output. What Huncher does with your data is narrower. Account details and authentication run through its auth provider; the URLs you submit and the public page content it crawls are processed to produce the analysis; basic logs, meaning timestamps, error traces and coarse device information, are kept to operate and secure the product. Crawled content is sent to an AI sub-processor solely to generate your results. One clause deserves a close reading: the policy says Huncher does not use your scan content to train third-party models beyond what producing your results requires, and says nothing at all about training models of its own.
Data retention & training
Hosting summary
Huncher names no hosting country and no hosting region anywhere on its site, so the jurisdiction your data sits in cannot be determined from what is published. What it does name is the infrastructure behind the service: Supabase for the database and authentication, Vercel for hosting and the sandboxed crawl environment, and Anthropic for the models that analyse pages and write fix prompts. The privacy policy adds that these providers may process data in regions outside your own, and that transfers are covered by appropriate safeguards such as standard contractual clauses where required, which is a statement of mechanism rather than of location. On security, the policy claims industry-standard measures, namely encryption in transit, scoped access controls and per-user data isolation, together with a commitment to notify material breaches as the law requires. No certification is claimed and no audit is referenced. One observation drawn from the technical record rather than the policy: the site itself is served from Vercel infrastructure, which is consistent with the disclosure.
Things to keep in mind
Risks and trade-offs to weigh before adopting Huncher.
- You cannot know what the tool costs before opening an account, which makes budgeting impossible and invites a commitment made in the dark
- The publisher is a brand with no legal identity, no address and an unnamed jurisdiction: if something goes wrong, there is no one identified to pursue and no court named to pursue them in
- Your crawled page content leaves for a third-party AI sub-processor, and the no-training promise covers third-party models beyond what generating your results requires while staying silent on Huncher's own
- Liability is capped at what you paid over the previous twelve months and the service is supplied as is, so a fix prompt that breaks a page is your problem alone
- The prompts edit your real repository: running one without reading it, on a codebase you have not backed up, is the fastest way to turn an audit into an outage
- Scores out of 100 and leak counts are persuasive by design, and treating them as a target risks optimising for the tool's checklist instead of for your actual visitors
- You are responsible for having the right to scan any URL you submit, robots directives included, an obligation easily forgotten when a competitor's site is one paste away
Setup & Integrations
Technical difficulty
Getting a diagnosis requires nothing technical: create an account and paste a URL. There is no script to install on the site being analysed and no repository access is needed to run the scan. Applying the fixes sits at a different level. The prompts are written for an AI coding agent operating on your codebase, so you need to be comfortable running one and reviewing what it changes. The MCP route, which means wiring Huncher into Cursor, Claude or v0, assumes you can configure an MCP server, and that is developer territory. Easy to try, developer-grade to act on.
Deployment
Integrations
Behind Huncher
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Huncher actually do?
How many pages does a scan cover?
How do I apply the fixes?
Which AI tools does it work with?
What does it cost?
Can I scan any website I want?
Is there a minimum age?
Who processes my data?
Is my scan content used to train AI models?
Is there a mobile app?
Should you pick Huncher?
Huncher answers a narrow question well. Not what is wrong with this site, which plenty of tools will tell you, but what exactly do I hand my AI coding agent to fix it. The crawl, the ICE ranking and the four-way read are competent without being unusual; the executable prompt at the end of each finding, and the MCP server that removes even the copy-paste step, are what set the product apart. For a team already building in Cursor, Lovable or v0, that closes a loop most audit tools leave open.
The reservations are real, and they concern the publisher more than the product. Nothing on the site identifies a legal entity: no company form, no postal address, and a governing-law clause pointing to an unnamed operating entity whose courts would hear any dispute. No price is published, so the cost of the second page is unknowable until you hold an account. There is no DPA, no security certification and no named hosting jurisdiction, which is enough on its own to stop a procurement process. The domain was registered on 13 July 2026, making this a product with weeks of public history, no cited customers and no social presence.
None of that makes Huncher a bad tool; it makes it an early one. The sensible way in is the one the site itself offers: scan the first page for free, judge the leaks it finds and the prompts it writes against your own site, and let that decide whether the missing commercial and legal transparency is a price worth paying for the output. Anyone handling client sites, or working under a procurement policy, should wait for that transparency before committing.
- Choosing a selection results in a full page refresh.
- Opens in a new window.