Skip to content
missumai
Steal this idea· 8 min read

Steal this idea: VectoSolve gives every file conversion its own landing page

A former landscaper built a PNG-to-SVG tool with Manus and Claude Code, then grew it almost entirely through search: one page for every “this format to that format” people look for, each with the converter on it. Here's the pattern, the proof, and the traps.

monthly recurring revenue, Stripe-verified, from 263 active subscriptions (11 Nov 2026)
$1,648
in the last 30 days, Stripe-verified, because credit packs sit on top of the subscriptions (11 Nov 2026)
$2,737
/convert/ landing pages in today's sitemap, 70 in English, by our count (11 Nov 2026)
299
users about two weeks after the first big batch of pages, founder-reported
10x

VectoSolve turns PNG, JPG and WEBP images into SVG vectors and into the files cutting, laser and embroidery machines read, such as DXF, G-code, DST, PES and JEF.⁠[1] It is run by Rob, a one-person team in France according to TrustMRR.⁠[1] He told AI Bites he started work in landscaping at 16, did plasterboard after that, ran a small digital agency, and at the end of 2025 launched a batch of small SaaS products, VectoSolve among them.⁠[2] His X bio describes a quest to build 1,000 SaaS products.⁠[11]

How he built it is short: he says he gave Manus his stack and asked it for a new micro-SaaS about 20 times, kept the ones worth keeping, and now writes each idea up as a Markdown file and builds it in a back-and-forth with Claude Code.⁠[2] What is worth stealing is how he got it found.

The pattern in one sentence

If people search for your tool by naming the job ("PNG to SVG", "SVG to PES", "image to G-code"), give every one of those jobs its own page, put the working tool at the top of each page, write what is genuinely different about that pair underneath, and add new pairs when customers ask for them.

What it has done so far

TrustMRR, which reads VectoSolve's Stripe account through an API key, showed $1,648 in monthly recurring revenue from 263 active subscriptions, $2,737 in the last 30 days and $23,207 all-time on 11 November 2026, with a founding date of October 2025.⁠[1] Three weeks earlier, Shipped & Paid read $2,695 in the last 30 days and $1,676 MRR from 253 subscriptions there.⁠[3] That is the verified part.

The rest is Rob's own account. On the founder-written part of the same TrustMRR page he lists 20,020 registered accounts and 1,892 paying customers in 98 countries.⁠[1] In the interview he puts his margin over 80% (the host's summary) and says about half of buyers take a one-time credit pack rather than a subscription, which is why monthly revenue runs well above MRR.⁠[2]

  1. Oct 2025

    Founding date on TrustMRR.⁠[1]
  2. End of 2025

    Launched with a batch of small SaaS products; early users were friends, then more visitors showing up in Search Console (founder-reported).⁠[2]
  3. Early Feb 2026

    Puts the site in every language and generates about 1,000 pages; about two weeks later he has 10 times the users, all from search (founder-reported).⁠[2]
  4. Feb 2026

    Lots of users, no money. He cuts free conversions from 25 to 3, and the first payment and first MRR arrive at the very end of February (founder-reported).⁠[2]
  5. Apr 2026

    MRR takes off, in his words, after daily SEO updates and customer feedback (founder-reported).⁠[2]
  6. 12 Oct 2026

    AI Bites interview.⁠[2]
  7. 11 Nov 2026

    TrustMRR: $1,648 MRR, 263 subscriptions, $2,737 in 30 days. Sitemap: 1,652 URLs, 299 of them conversion pages.⁠[1]⁠[4]

How the search engine was built

Rob describes the steps simply. First he looked at the competitors and benchmarked their keywords. Then he asked his AI to code a big SEO structure: the site in 12 languages and landing pages for PNG to SVG, then JPEG to SVG, then every export the tool could do.⁠[2] He says Manus was useful for the research side, scraping for the most used keywords, and that when he has an idea now he goes straight to Claude Code.⁠[2]

You can see the result in the sitemap. We counted 299 URLs under /convert/ on 11 November 2026: 70 English pages and 229 translated ones, with German (46), Spanish (36) and French (31) the largest.⁠[4] The English list reads like a search-term list: png-to-svg, jpg-to-svg, logo-to-vector, png-to-svg-for-silhouette, image-to-dxf, image-to-gcode, svg-to-pes, svg-to-dst and so on.⁠[4] Alongside them sit 287 English blog posts and their translations.⁠[4] Rob mentioned about 200 articles in the interview; the sitemap now lists more than that.⁠[2]

The tool is on every page. The SVG to PES page opens with "Upload your image to convert to PES", a drop zone and a free preview with no signup.⁠[7] A visitor who searched for that exact job can do it without clicking anywhere else.

Each page says what only that pair needs. The PES page explains that Brother's PES format carries a thread palette while Tajima DST does not, that Babylock and several Bernina machines read it, and how, by default, the design is fitted to a 4x4 inch hoop before any stitch is generated.⁠[7] The G-code page shows a real conversion of a scan from a 1914 book, made with the page's own settings, with the number of cut paths and the millimetres of cutting.⁠[8] The PNG to SVG page explains what an SVG is, when not to use one (photos), and how the file behaves in Cricut Design Space.⁠[6] The pages also carry a visible "Last updated" date: 2 May, 11 September and 3 November 2026 on those three.⁠[6]⁠[7]⁠[8]

“And when the design cannot be stitched, no file is produced and you are not charged.”
VectoSolve's SVG to PES page⁠[7]

Customers add the pages. Rob says he learned that many visitors didn't just want PNG to SVG; they wanted files for a specific machine, so CAD, CNC and other exports came from their requests, each with a new landing page.⁠[2] He puts it at about 80% of features coming from what customers ask for.⁠[2]

Backlinks, the boring way. Early launches on PeerPush, Product Hunt and Reddit brought the first links; about two months later he emailed blogs in his niche to swap links.⁠[2]

The part that made it pay

Traffic alone didn't do it. Rob says he had 10 times the users all through February and made no money, until he cut the free tier from 25 conversions to 3; the first payments followed.⁠[2] Today the free option is smaller still: a watermarked preview with no signup, one a day according to the founder's TrustMRR text, with a single clean file for $6.99 or plans from $7.99 a month for 25 conversions.⁠[1]⁠[10]

That matters for anyone copying the pattern. A page per search term brings people who want one file, once. A cheap one-off unlock next to a subscription lets them pay instead of leaving, and Rob says about half of his buyers do exactly that.⁠[2]

The stack, as described

Built withManus to generate the first batch of products; Claude Code for day-to-day building from a Markdown spec.⁠[2]
App and hostingNext.js on Vercel, with Supabase.⁠[2]⁠[1]
PaymentsStripe, subscriptions plus one-time credit packs.⁠[1]⁠[10]
The conversionHe says he first used Recraft, and now runs an open-source model plus JavaScript scripts, which keeps costs down.⁠[2]
SupportEmail and a Crisp chat widget.⁠[2]
Other ways inA Chrome extension, a Firefox add-on, an embeddable widget, a REST API, a hosted MCP server and more, per the founder's TrustMRR text.⁠[1]

How to build your own

The steps follow what VectoSolve did. Where a step is our own suggestion, it says so.

  1. Pick a tool people search for by naming the job

    File converters are the obvious case: the query is the product. Rob started with a problem of his own, turning lots of small logos into vectors.⁠[2] The same shape fits unit converters, calculators or "X template for Y" tools (our suggestion).
  2. List the pairs before you write a page

    Look at what competitors rank for and benchmark the keywords first, as he did.⁠[2] Then write every input, every output, and the audience words people add ("for Cricut", "for Silhouette").⁠[4]
  3. Make one template, then make every page earn its place

    The shared parts (upload box, steps, FAQ) can be a template; the section that explains the pair should not be. VectoSolve's PES, G-code and PNG pages each explain something the others don't.⁠[7]⁠[8]⁠[6]
  4. Put the working tool above the fold

    Upload, preview, then pay, on the same page the search landed on.⁠[7]
  5. Show a real result

    A worked example made with the page's own settings, with its measurements, is proof no competitor can copy-paste.⁠[8]
  6. Translate only what earns traffic

    Rob went to 12 languages early, but says that after a big Google update you may not need to now.⁠[2] The site's robots.txt notes say languages with no measured Search Console traffic are blocked, while ones that brought clicks keep redirecting to English.⁠[5]
  7. Get a few honest links

    Launch directories and Reddit first, then trade links with blogs in your niche.⁠[2]
  8. Price for the one-file visitor

    Keep the free taste small and add a single-file unlock next to the plans.⁠[2]⁠[10]
  9. Let support tickets write your next pages

    New formats came from customer requests, and each one got a page.⁠[2]

How to stand out if you copy it

  • Choose a niche where the pairs really differ. Machine formats are a good example: Rob says Cricut, laser and embroidery each need a different file, and that it gets complex fast.⁠[2] A page that explains the difference is useful; a page that swaps one word is not.
  • State your limits. The PES page lists what the converter does not do and refuses files it can't stitch rather than selling a broken one.⁠[7] In our view that builds trust and saves support time.
  • Own one audience. VectoSolve's founder text names crafters and makers first, then print and e-commerce sellers.⁠[1] Write every page for that person, not for "everyone who has a PNG" (our suggestion).

Risks before you copy it

  • Google has a name for doing this badly. Its spam policies say: "Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users."⁠[12] Generating pages with AI is listed as an example only when they add no value, so the per-pair detail is what keeps you on the right side.⁠[12]
  • Half-translated pages. On the German PNG to SVG page we opened, the title and intro are German while the comparison section below stays in English.⁠[9] Translate fewer pages well rather than many pages halfway (our suggestion).
  • Traffic is not revenue. A month of ten times the users and no payments, until the free tier shrank.⁠[2]
  • Search moves under you. Rob himself says a Google update changed what works.⁠[2] One channel is one dependency.
  • One-file buyers are hard to reach again. He says about half of customers don't want email, and that around 5,000 emails got him roughly 10 replies.⁠[2]
  • Comparison tables go stale. The converter pages compare VectoSolve with other tools using "Data as of March 2026".⁠[6] If you name competitors and their prices, date it and recheck it.

What to take from this

  1. 01

    The search term is the page

    One job, one URL, the tool on top.⁠[7]
  2. 02

    Specific beats many

    Each pair page explains what only that pair needs.⁠[7]⁠[8]
  3. 03

    Customers pick the next page

    New formats came from requests.⁠[2]
  4. 04

    Fix the free tier early

    25 free conversions brought users; 3 brought payments.⁠[2]
  5. 05

    Sell the single file too

    About half of buyers pay once.⁠[2]
  6. 06

    Prune by data

    Languages without measured traffic are switched off.⁠[5]

Search as the main channel shows up across smaller AI-built apps too; see AI-built apps making under $10K a month, where VectoSolve is one of the picks, and how GainFrame was built for a Claude Code-run SEO routine.

All stories

The newsletter

More stories like this, by email

Occasional email with new projects on the leaderboard and new build stories. Every number sourced, like the site.

We'll email a link to confirm. Unsubscribe in one click.

Sources

  1. [1]VECTOSOLVE on TrustMRRTrustMRR (revenue verified with a Stripe API key) · checked Nov 11, 2026Figures as shown on 2026-11-11 (the page says last updated that day); they move daily. The description, audience, user and customer counts, pricing and “additional info” on that page are the founder's own text.
  2. [2]$3K mrr AI SaaS founder did it in 6 months!AI Bites on YouTube (interview with Rob, VectoSolve's founder) · Oct 12, 2026 · checked Nov 11, 2026We read YouTube's automatic English captions. They are machine-made and garble names (“Vector Solve”, “cloud code”), so we paraphrase and quote nothing from them. Every figure from the interview is the founder's own.
  3. [3]VectoSolve: a PNG-to-SVG converter at $2,695 a month (Stripe-verified), built by a former landscaper with Claude Code and ManusShipped & Paid · checked Nov 11, 2026A summary of the AI Bites interview, with TrustMRR figures read on 2026-10-20.
  4. [4]VectoSolve sitemap index and its four sitemapsVectoSolve · checked Nov 11, 2026We downloaded /sitemap.xml and /sitemap/0.xml to /sitemap/3.xml on 2026-11-11 and counted the URLs ourselves: 1,652 unique URLs, 299 of them under /convert/ (70 in English, 229 in ten other languages).
  5. [5]VectoSolve robots.txtVectoSolve · checked Nov 11, 2026Read on 2026-11-11. The comments in the file are the site's own notes on its crawl and language decisions.
  6. [6]PNG to SVG Converter — Free AI Vectorizer, No SignupVectoSolve · checked Nov 11, 2026Read on 2026-11-11. Pages on a live product change often; the page shows its own “Last updated” date.
  7. [7]SVG to PES Converter — Brother Embroidery 2026VectoSolve · checked Nov 11, 2026Read on 2026-11-11. Pages on a live product change often; the page shows its own “Last updated” date.
  8. [8]Image to G-code Converter 2026VectoSolve · checked Nov 11, 2026Read on 2026-11-11. Pages on a live product change often; the page shows its own “Last updated” date.
  9. [9]PNG zu SVG Konverter — Kostenlos, KI-VektorisiererVectoSolve · checked Nov 11, 2026Read on 2026-11-11. Pages on a live product change often; the page shows its own “Last updated” date.
  10. [10]Pricing: VectoSolve | AI Image to SVG Plans from $7.99VectoSolve · checked Nov 11, 2026Read on 2026-11-11. Prices and limits change; check the page before you compare.
  11. [11]Rob☕️ (@Go_To_Rob) on XX · checked Nov 11, 2026Profile bio read on 2026-11-11 through the public fxtwitter API.
  12. [12]Spam Policies for Google Web Search: scaled content abuseGoogle Search Central · checked Nov 11, 2026

Something wrong or out of date? Tell us and we'll fix it.