The Landing Page Is Dead; Long Live the Demo
If you run growth or social for a software product, you already know the dirty secret of the funnel: the click is easy, the install is hard. Every link you post on LinkedIn, every carousel you ship on Instagram, every “link in bio” you rotate on TikTok ultimately dumps a stranger onto a static marketing page that can describe your product but can never let them touch it. That gap — between the promise and the proof — is where most creator-led and founder-led funnels quietly die. So when a tool shows up claiming to turn any link into the actual product, running in a cloud computer, no install required, my ears prick up. That’s Portal, and it just opened its self-serve builder.
What Portal Actually Is (And Why It’s Not Just Another Demo Tool)
The maker, Zachary Goldman, frames the thesis cleanly: a link can reach a marketing page, but it can’t reach the product. Portal’s answer is to put the real app inside a cloud computer and hand you a URL. Whoever opens it lands inside a live, stateful environment you’ve pre-configured — signed into a demo account, with buttons or API calls optionally blocked via a no-code builder, and with an optional AI that can see the screen and answer questions while the visitor pokes around. You can be in there with them, or they can explore alone.
That last detail is the part most people skim past, and it’s the part that matters most for social operators. The maker’s own use cases read like a creator’s wishlist: sending a Portal link during a FaceTime call instead of screen-sharing so both people can click; learning a tool like Photoshop with an AI tutor; dropping research participants into a safe account to reduce friction. The listed user patterns include autonomous user research, first-round interviews where a candidate performs a task in a real account, intake conversations, using it as a business-card replacement, whiteboarding with teams mid-call, and support onboarding for enterprise trials that can’t afford 1:1 calls for every prospect — set up a demo account once, use Portal’s reset hook, and let prospects make a mess safely that gets cleaned up in the database. For forward-deployed engineers, it’s a way to send a proof of concept instead of asking a client to sign up cold.
The company is backed by South Park Commons, and the self-serve builder is the new piece — previously you’d have needed to work with the team. Pricing for the self-serve tier is not disclosed in the launch material. The demo link points at a Portal signed into Linear in a demo account, and you can make one for free at makeportals.com.
Why this is a social-media story, not just a SaaS story
Here’s my take: the creator economy has spent five years optimizing the top of the funnel — hooks, thumbnails, retention curves, comment-bait — and almost no energy on the bottom. We’ve built an entire discipline around getting someone to tap a link, then handed them a landing page that asks them to book a call. Portal is interesting because it attacks the exact moment where social traffic converts or evaporates. If you’re a growth marketer running paid social for a B2B tool, or an indie founder whose entire distribution is a YouTube channel and an X account, the difference between “watch my demo video” and “click here and actually use it” is the difference between a lead and a user.
How It Stacks Up Against the Incumbents You Already Pay For
Let’s be honest about the competitive set, because “interactive demo” is not a new category.
The closest well-known incumbent is Navattic, which builds no-code interactive product tours — screenshots stitched into clickable flows. Storylane does something similar, and Arcade has carved out a niche with its Chrome-extension capture workflow. These tools are genuinely good at what they do, and they’re cheaper and lighter than spinning up a cloud computer per visitor. But — and this is the crucial distinction — they are approximations. They’re replicas of your product, not your product. The maker explicitly calls this out: “build screen recordings, clickthroughs, replicas, and sales calls — all approximations of the thing itself.”
Then there’s the screen-share-and-video tier: Loom, Vidyard, and the native screen recording in CapCut or Canva if you’re doing it for social. These are cheap, fast, and they scale across TikTok, Reels, and YouTube Shorts. But a video is a monologue. The viewer can’t click. They can’t break it. They can’t discover the feature you forgot to mention.
And then the heavyweight: FullStory, Heap, and the analytics layer. Portal’s own analytics were praised by a customer in the launch comments — Rahul Arulkumaran, who runs Browzer, said he’d been a customer and called it “an awesome screening tool to gauge interest from potential customers,” specifically flagging the analytics features. That’s a real signal, because the analytics are where the ROI argument lives: if you can see which prospects actually clicked around inside your product versus which ones bounced off the landing page, you’ve got a lead-scoring signal that no video view can give you.
Where the math breaks
Here’s my skepticism, and I want to be precise about it. A cloud computer per visitor is expensive. Video hosting is nearly free at scale; a Navattic tour is essentially a static asset. Portal is running live instances. The launch material doesn’t disclose pricing, instance limits, concurrency caps, or session timeouts — all of which determine whether this is viable for a creator with 50,000 followers who posts a link and gets 3,000 clicks in an hour. If the economics only work for enterprise sales teams doing 20 demos a month, then the “creator” framing is aspirational, not operational. I’d want to see the actual session limits before I told a client to swap their landing page for a Portal link.
There’s also the security and abuse surface. You’re handing anonymous internet traffic a live environment. The no-code builder that blocks buttons and API calls is the right instinct, and the reset hook is smart, but the failure mode — someone finds a path you didn’t block — is worse than a broken demo page. The maker’s own origin story is instructive here: he was running user research and people wouldn’t install the extension because they were afraid it was malware. The irony is that Portal asks for a different kind of trust — you’re clicking a stranger’s link and landing inside a live system. That’s a higher-trust ask, not a lower one, and it needs to be communicated carefully in any social post that promotes it.
What Creators and Social Teams Can Actually Steal From This
Even if you never sign up for Portal, the pattern is worth stealing, because it maps onto problems every social operator has.
Why TikTok creators should care more than LinkedIn ones
The obvious Portal use case is B2B: enterprise trials, FDE proof-of-concepts, user research. That’s a LinkedIn and cold-email story. But the more interesting angle is consumer-facing. Think about what a TikTok creator selling a digital product — a Notion template, a Lightroom preset pack, a spreadsheet system — actually sends people to. A Gumroad page with screenshots. A Stan Store link. A preview image. The buyer can’t use the thing before buying. If you could send a link that drops them into a live, sandboxed version of your template with an AI answering “what does this column do?” — that’s a fundamentally different conversion event. I’d bet the first wave of creator adoption comes from template and tool sellers, not enterprise sales teams, precisely because the trust barrier is lower and the “try before you buy” instinct is already normalized on TikTok.
The repurposing angle nobody’s talking about
Here’s an operational detail I keep coming back to. The maker mentions using Portal to learn Photoshop with an AI tutor, and to FaceTime his parents with a shared clickable screen. Both of those are content formats. A “let’s learn this tool together” video where the AI tutor is answering live questions is a YouTube format that doesn’t exist yet at scale. A “come break my product” livestream where viewers click around a Portal link and you watch them in real time is a Twitch or Threads native format. The tool is a demo tool, but the content it enables is closer to interactive television than to a SaaS onboarding flow. If I were running social for a dev tool right now, I’d be testing a weekly “Portal office hours” stream where the audience gets a link and tries to break the thing on camera. That’s watch time, that’s engagement rate, that’s a clip factory for Instagram and TikTok.
The UTM and attribution problem
One thing the launch material doesn’t address, and which matters enormously for social operators: attribution. If you post a Portal link on five platforms, can you tag it with UTMs? Does the analytics layer distinguish a Pinterest visitor from a Facebook visitor? Does it integrate with your CRM or your HubSpot pipeline? For a creator running a multi-platform funnel, a demo link that doesn’t pass attribution data back to your analytics stack is a black box. I don’t know the answer here — it’s not disclosed — but it’s the first question I’d ask before making Portal the primary CTA in a campaign. A tool that generates leads you can’t attribute is a tool that gets cut in the next budget review.
Where My Judgment Says It Falls Short
Let me be balanced, because the launch page is a launch page and I’ve been burned by launch pages before.
The self-serve builder is brand new. The maker’s own framing — “now launching the self-serve builder” — tells you the product has been high-touch until now. Self-serve tools live and die on onboarding friction. If setting up a Portal requires you to understand your product’s API surface, block the right endpoints, and configure a reset hook, then the “no-code builder” is doing a lot of heavy lifting in that sentence. I’d want to see how long it takes a non-technical social manager to go from signup to a shareable link. My guess is the first 20% of setup is easy and the last 20% — the part that keeps visitors from breaking things — is where people churn.
The AI tutor is a feature, not a moat. An AI that “sees the screen and answers questions” is increasingly table stakes. OpenAI, Anthropic, and every screen-sharing tool are converging on this. Portal’s edge, if it has one, is the environment — the fact that the AI has a live product to reason about — not the model itself. That’s a real edge, but it’s an edge that requires the cloud-computer infrastructure to stay cheap, which brings us back to the economics question.
Who this is NOT for. If your product is a mobile app, a Chrome extension, or anything that requires a native install, Portal’s cloud-computer model doesn’t obviously map. If your funnel is high-volume, low-intent top-of-funnel social traffic, the cost-per-session math probably doesn’t work. And if your product is simple enough that a 30-second Loom video explains it, you don’t need this. Portal is for products that are hard to explain and easy to understand once you’re inside — which is a narrower category than the launch page implies.
The trust framing is underdeveloped. The origin story is about malware fear, but the product asks for a bigger leap of faith. A social post that says “click this link and you’ll be inside a live version of our product” reads, to a skeptical scroller, like a phishing attempt. The marketing needs to do the trust work that the product can’t do on its own. That’s a content problem, and it’s the kind of thing a good social operator solves with a 15-second explainer video, not a landing page.
What I’d Watch / Test Next
If you’re an operator reading this, here’s what I’d actually do this week, in order.
First, go make a free Portal at makeportals.com and time yourself. Not the signup — the setup. How long until you have a link you’d be comfortable posting publicly? That number tells you whether this is a tool or a project.
Second, look at your current link-in-bio or primary CTA across Instagram, TikTok, and LinkedIn. If it’s a landing page with a “book a demo” button, run a two-week split: Portal link in bio versus landing page in bio, same content, same posting cadence. Watch the click-to-engagement rate, not the click-through rate. The whole thesis is that engagement beats clicks.
Third, check the attribution question before you scale. Ask whether Portal passes UTM parameters and whether the analytics export to wherever you track pipeline. If it doesn’t, treat it as a top-of-funnel experiment, not a core funnel component.
Fourth, watch the pricing page. The launch material doesn’t disclose it, and the entire creator-market thesis hinges on whether a solo operator can afford hundreds of sessions a month. If the self-serve tier prices like a Navattic or Storylane — enterprise seats, annual contracts — then this stays a sales-team tool and the creator angle is marketing. If it prices like a Loom — per-seat, monthly, cancel anytime — then it’s a genuine new primitive for social funnels.
My take: the idea is stronger than the current product’s likely execution, and that’s fine. Ideas like this get copied fast. The question isn’t whether Portal wins — it’s whether “the link goes to the product, not the pitch” becomes the default expectation for software marketing. If it does, every social operator reading this will eventually have to rebuild their funnel around it. I’d rather test it now, while it’s cheap and weird, than retrofit it later when it’s obvious.
