VellumUp

Getting positive feedback is obviously great, but what has helped the most are the suggestions, bug reports, and small UX comments people have taken the time to share.

At first, I was honestly struggling to keep track of everything. Feedback was coming from different places, conversations were scattered, and I kept worrying I'd miss something important.

So I ended up building a small internal system to organize feedback, track requests, and make sure nothing falls through the cracks. Nothing fancy, just something that helps me stay on top of things.

Looking back, a huge portion of the improvements we've made to VellumUp came directly from user feedback. The product is significantly better today, both in terms of UX and overall quality, than it was when we launched.

It's been really rewarding to see the product evolve so quickly, and honestly, it's a great feeling to wake up, improve something based on real feedback, and keep moving forward.

Still a long way to go, but I'm excited about where it's heading.

If you're curious, feel free to take a look at where we've gotten so far: vellumup.com

And if you have feedback, suggestions, or things you'd improve, I'd genuinely love to hear them.

Hey everyone,

I launched my SaaS, VellumUp, not too long ago, and I just wanted to share a small win that has me really pumped today.

Over the last 3.5 days alone, 14 new users signed up. I know that in the grand scheme of things, 14 users might look like a drop in the ocean. But when you build something from scratch, watching that user count steadily tick up day after day is honestly the best feeling in the world.

For a bit of context: VellumUp is an SEO platform. I built it because I got tired of generic AI tools that just output walls of text. It deeply analyzes your website to understand your brand voice and product angle, and then automatically creates and publishes tailored articles designed to rank on Google and more importantly nowadays get cited inside AI search engines like ChatGPT and Perplexity.

Seeing people actually log in, connect their sites, and start using it to grow their organic traffic makes all those late-night coding sessions worth it.

If anyone is struggling with getting their product noticed by AI search engines, or just wants to check it out and give some feedback, I'd love to chat. You can play around with it at vellumup.com (you get 5 free articles to test the waters, no strings attached.).

Trying to automate SEO content always felt like choosing between speed and actually being proud of what goes live.

The trigger for building vellumup was simple: I wanted everything around SEO, GEO and even AEO to be automatic, but I refused to accept the usual drop in quality or that generic AI tone everyone recognizes in 2 seconds. The content had to still sound like me, actually know my site, and keep up with constant changes so it did not become stale.

Here is what I ended up focusing on:

- I treated "automatic" as non‑negotiable, but only if the content stayed at the same level I would write myself.

- The system had to learn the real voice from the existing site, not from a few prompts or one template.

- It needed to keep updating in the background, for example when internal links change or when something shifts in the SEO or GEO world.

- Internal structure mattered as much as the words, so I wanted it to notice and react to things like new pages that should be linked together.

- The core idea became: automation should feel like an extension of you, not a replacement for you.

Right now on vellumup.com I kept things very low friction: free access with no commitment so people can generate up to 5 articles and see how it behaves on their own site.

What surprised me most was how hard it is mentally to fully trust automation when it touches something as personal as your writing voice. The technology part was one thing, but letting it actually publish in a way that still feels like "you" was the real challenge.

Has anyone else tried to push SEO automation to the max without accepting that generic AI tone, and what tradeoff hit you first?

I built VellumUp.com after noticing that most AI SEO tools generate content without truly understanding the business they're writing for.

The process starts with a simple website URL.

VellumUp.com performs a deep analysis of the business, its products, services, positioning, brand voice, site structure, and existing content. It builds a detailed profile of the company and uses that knowledge to create a content strategy tailored specifically to that business.

The platform researches keyword opportunities, identifies content gaps, creates a monthly content plan, writes articles aligned with the brand, adds relevant internal links, and publishes directly to the website.

What makes it different is that it continuously learns and adapts. As the website grows, new pages are added, products change, and search behavior evolves, VellumUp.com updates its understanding of the business and adjusts its strategy accordingly.

The goal is simple: help business owners grow their visibility on Google and AI search without spending hours on SEO, content planning, or content creation.

We're building for the future of search, where businesses need visibility not only in traditional search results but also in AI-powered search experiences and answer engines.

We've recently launched and are starting to get our first users.

I'd love feedback from other indie hackers on the positioning, onboarding experience, and overall value proposition.

We're offering free access with no commitment required so new users can generate up to 5 articles for free and see how the platform works on their own website.

Feel free to try it yourself at VellumUp.com.

In 2026, the biggest shift in search is not "AI answers instead of blue links". It is that authority has moved from the page level to the claim level.

In classic SEO, the mental model was: rank pages. If your page had enough links, relevance, and on-page tuning, Google rewarded it as a unit. The whole URL rose or fell together.

In GEO, generative engines like ChatGPT, Perplexity, and Gemini do something different. They read your page, slice it into small passages, turn those into embeddings, and then cross check those passages against thousands of other sources. When they answer a prompt, they are not thinking "Which page should I rank?". They are thinking "Which specific claims are safe and useful to repeat?".

That subtle change explains a lot of the chaos founders are seeing this year:

- Your old SEO winners are invisible in AI summaries even though they still rank in Google

- A tiny blog post can suddenly become a citation magnet for certain prompts

- Traffic from AI assistants feels spiky and hard to predict

The model does not really trust your page. It trusts (or ignores) individual claims.

Once you accept that, GEO vs SEO stops being mysterious and turns into a content design problem.

Here is the working model I use for my own projects:

1. Every serious page is a bundle of claims

Facts, instructions, comparisons, benchmarks, opinions. Each one is an atomic piece the model can quote, paraphrase, or discard.

2. Claims are scored on three things

- Clarity: Can a model easily extract the statement and understand what it asserts?

- Support: Is there evidence, method, or at least consistent corroboration across the web?

- Context: Does the surrounding text explain when this claim applies and when it does not?

3. GEO engines pull from the top of that internal "claim leaderboard"

They select the cleanest, best supported, best contextualized claims across all sources, then weave those into an answer.

If that mental model is roughly correct, a few practical consequences fall out for founders.

First, vague content dies. The old SEO trick of writing 2,000 words of generic text around a keyword is not just useless now, it is reputational debt. A model parses that and finds almost no precise, checkable, differentiated claims. So your content becomes noise in its internal index.

Second, small sites can win narrow topics. If you publish one sharp, concrete, well supported claim about a specific question, you can punch far above your domain authority. For example, a tiny B2B tool that shares real, method backed data about "average onboarding time for X" can become the default citation for that fact.

Third, contradictions hurt more than you think. If page A on your site says "our average response time is 200ms" and page B says "our average response time is 400ms", a model trying to answer "How fast is X?" now treats you as a fuzzy source. Claim level authority means internal consistency matters more.

So what do you actually do with this as an indie founder?

1. Rewrite around atomic claims

Go through your key pages and highlight specific statements that a model could quote. Are they clear, specific, and free of weasel words? Can you restate them in one clean sentence without losing meaning?

2. Attach method and evidence

Right next to each important claim, add how you know. Sample size, time window, data source, even a short note like "based on 327 user sessions from Jan to Mar 2026". Models are getting better at spotting this and ranking those claims higher.

3. Make claims machine friendly

Use tables, bullet lists, simple subheadings, and consistent phrasing. Mark up key data with schema where it makes sense. You are not just writing for humans anymore, you are writing for vector indexes.

4. Audit contradictions across your own site

Keep a simple "claims registry" in a doc or code repo. For each key metric or positioning statement, track the current canonical value and where it appears. When you change it, update all surfaces.

The frustration that pushed me to build VellumUp was exactly this claim level reality. I watched my own long form posts get ignored by AI engines while one or two precise, well supported paragraphs quietly drove all the assistant mentions. It forced me to stop thinking in "articles" and start thinking in "claims inventory".

SEO used to ask: "How do I get this page to rank?".

GEO now asks: "Which claims on the internet deserve to be repeated?".

Founders who adapt to that question early will own more of the answer space for their niche, even if their domains are tiny.

How are you currently structuring your content to create trustworthy, quotable claims that AI assistants will actually pick up in 2026?

If you're trying to get your site seen by both Google and AI search engines without spending hours on content every week, that's exactly the problem I built vellumup.com to solve. Happy to answer questions in the comments.