Answered by the same rule: only what we can prove.

Six scored categories: UX and conversion, performance, technical SEO, accessibility, security and privacy, and content and trust. The quick audit reads the homepage, robots.txt, the sitemap and a handful of same-origin links. In the background it adds Google PageSpeed lab runs on mobile and desktop and a sealed browser capture at 1440 and 390 pixels wide, which measures layout, legibility, tap targets, images and contact routes.

The quick audit shares a 3.5-second budget, so a report usually appears within a few seconds. PageSpeed and the rendered capture keep running afterwards and update the same report — you’ll see which measurements are still pending.

It fails safely. An unreachable homepage produces no invented findings. Blocked, rate-limited and timed-out requests stay unverified, and a link only counts as broken when a real request confirms a 404 or 410.

No — on purpose. Each audit is bounded: up to 1 MB of homepage HTML, a few internal links and, for contact discovery, a small number of contact, about and imprint pages. Bounded checks are fast, polite to the website and honest about what they did not see.

Confirmed findings were directly observed, such as a missing tel: link or a measured lab timing. Heuristic findings are likely but depend on thresholds or context. Manual-review items are things a bounded check cannot settle, such as the absence of a contact page. Manual-review items never reduce a score and never reach an email.

No. MovaScope is a bounded review of public signals. It isn’t legal advice, a penetration test, an accessibility certification or proof of compliance — and it never presents itself as one to your prospects.

Each category starts at 100. Findings deduct 45, 30, 18 or 5 points for critical, high, medium or low severity, scaled by confidence (100, 80 or 55 percent). Categories are weighted — UX and conversion 30 %, performance 20 %, technical SEO 15 %, accessibility 15 %, security and privacy 10 %, content and trust 10 % — and the total blends 85 % of that weighted mean with 15 % of the weakest visitor-facing category.

Because something hasn’t been measured yet. Unmeasured categories are left out and the remaining weights renormalised, instead of counting them as zero or crediting them with 100. Once PageSpeed and the visual capture complete, the score can change and the label disappears.

A separate ranking of who to contact first. It combines redesign need (45 %), commercial signals such as reviews, rating, industry and location (30 %), contact confidence (15 %) and the website-quality gap (10 %). Every number comes with plain-language reasons. Unknown information is never treated as negative.

Background measurements update the report they belong to. Reports keep the algorithm they were created with; re-run an analysis to apply the latest scoring.

No. First emails are composed deterministically from your saved agency profile, the company’s context and one verified finding, in English or German. There is no model in the email path, so there is nothing to hallucinate.

Only findings that are confirmed, high-confidence, carry evidence and need no manual verification. The strongest owner-relevant one is used; a second, independent finding is added only when the website scores below 65. Raw scanner wording never reaches the copy.

Then the draft doesn’t pretend. It uses a neutral introduction instead of alleging a problem that wasn’t observed.

Drafts need your explicit review and recipient confirmation, and editing a draft clears that review. Scheduled automations can stop at “Pause for my review”; they only send on their own when you set them up to.

Resend or Brevo, from your own verified domain. Credentials are encrypted before storage and never sent back to the browser. A provider’s acceptance is labelled as accepted, not delivered.

To your own reply-to inbox. Connect it over IMAP and replies to campaign emails are imported into MovaScope’s Inbox, next to Sent and Drafts.

If you connect native open tracking from Resend or Brevo, opens are recorded as “Open detected” — never as “Read”. Known privacy-proxy opens are counted separately.

MovaScope reserves companies and recipients across campaigns and gives every draft a single idempotency key, so retries and overlapping campaigns don’t produce duplicate outreach.

A redesigned homepage hero in the prospect’s own brand. MovaScope ranks twenty-five curated design systems for the company, pulls its logo, colours and photography from the rendered site, keeps strong copy and rewrites weak copy in response to the findings.

No — and the outreach never says it is. It’s a concept of the homepage’s opening section, offered as a visual idea rather than a free website.

Only people you share it with. Links are private, stored only as a hash, marked noindex and revocable; a newer concept supersedes the old one. Previews contain no live form submission.

Optionally, for art direction only. A model may propose a schema-constrained creative blueprint; every claim and evidence reference is validated, and anything that fails falls back to the deterministic direction. The model never receives raw website HTML as instructions and cannot add code.

Yes. Each concept is versioned, you can switch design systems for a new version, and it can be exported as a portable ZIP.

From real business listings in fifty curated markets. Places without a website, permanently closed listings, directories, social profiles and public institutions are excluded, and shared domains are deduplicated.

Only once its audit succeeds and a usable email address is found. Targets of 10, 20, 50 or 100 are never padded with failed or email-less companies; if a market runs dry, you’ll see the real count.

From published text and mailto links on the company’s own contact, about, team and imprint pages. Own-domain addresses take priority. Every contact keeps its source page and a confidence percentage, and stays labelled unverified.

Yes. Schedule up to twenty campaigns at once, each with its own date and start time in one of five time zones. Defaults are 20 emails per campaign, one minute apart, for websites scoring below 70.

Failed analyses and reused results are refunded. A successful audit that finds no email keeps its report and counts as a normal analysis.

No. Every read and write is scoped to your workspace on the server, with row-level security in the database as defence in depth.

Findings, their evidence and a bounded profile of public copy such as the title and headings. Raw crawled HTML and raw PageSpeed payloads are not stored.

Only public HTTP(S) addresses on standard ports are fetched. Credentials in URLs, localhost, private and reserved addresses and unsafe redirects are rejected, and every request has size and time limits.

Yes. Your workspace overview can be exported as JSON from Settings.

20 website analyses, once, with no credit card — plus one active campaign, opportunity scoring, evidence-gated drafts and human review before sending.

Volume. Plus includes 2,000 analyses a month and 10 active campaigns for €49 a month, or €39 a month billed annually. Ultra includes 5,000 analyses a month and unlimited campaigns for €119, or €99 billed annually.

Online checkout isn’t connected yet, so registration always opens a Free workspace. Paid plans are switched on for your account by the MovaScope team — get in touch and we’ll activate Plus or Ultra.

Paid allowances reset every month. The Free allowance is a one-time quota.