Free tiers that are actually usable, and the ones that are demos

We went into this expecting to write a list of watermarks and storage caps, and mostly we did. But the more useful finding was how unevenly "free" is applied across the fifteen tools we ran for a month of ordinary use — a personal page, a small content collection, a handful of real edits every few days. Some of these free plans are the product. A few are better than the cheapest paid tier on the same page. Most exist to be outgrown on a schedule the vendor chose, and a couple don't exist at all, whatever the pricing page implies.

The useful split isn't free versus paid. It's whether the ceiling sits above the work you actually do or below it, and if below, exactly where.

Three categories, and the threshold that defines each one

Genuinely sufficient means a single person doing real, ongoing work never touches the limit. Sufficient until a threshold means the free plan is real but has a specific point — a page count, a domain requirement, a storage figure — where it stops being usable and you either pay or stop. Unusable by design means the free tier was never meant to hold finished work; it's a preview with a shape that looks like a plan.

Webstudio's free tier sits in the first group cleanly: unlimited projects, unlimited pages, 250 MB storage, 300 form submissions a month, and — the detail that matters most — no badge at all. The only thing free withholds is a custom domain; everything else that makes a site a site is unrestricted. Google Sites is the other clean case: free forever on a personal Google account, no storage figure to watch because there isn't a separate one to hit, and no version of the paid product is actually aimed at an individual — Workspace plans start at $7 a month per user and exist for organizations, not solo users. And Canva Websites' free tier is unusually generous for the category: up to five live sites, 5 GB, 200 standard and 20 premium AI uses a month, all under xyz.my.canva.site.

The second group is where the interesting thresholds live. Wix's free plan is permanent and needs no card, but it caps at 500 MB storage and 1 GB bandwidth, and the Wix banner and favicon are compulsory — not a design choice you can turn off, a fixed condition of not paying. Webflow's Starter plan is the sharpest threshold in the whole set: two static pages. Not two megabytes, not two collections — two pages, full stop, before the badge-locked mysite.webflow.io subdomain is even worth publishing. Framer's free tier gives 500 one-off credits and 1 GB of bandwidth with a visible badge; the page limit isn't published, which is itself worth noting since every other number on that page is. Bubble's free tier runs on 50,000 Workload Units — third-party figures, not officially confirmed by Bubble, but the mechanism is documented: once you're building an actual app rather than a static page, that allowance moves fast.

The third group doesn't really have a free tier, whatever the homepage suggests. Squarespace says so outright — there is no free plan, only a 14-day trial, and the site stays private the entire time. Ghost is the same shape: no permanent free plan, a 14-day trial, and the honest free route is self-hosting the open-source software yourself, server costs and all. Butternut AI doesn't have a $0 plan either; "free" there means generating and editing drafts during a five-day trial, and whether a free site can be published publicly at all can't be determined from the vendor's own site, because two of its own pages contradict each other on the point.

The tactics that turn a free plan into a demo

Three mechanisms do almost all the work.

The first is the watermark, and it comes in two strengths. A cosmetic badge — Wix's banner, Carrd's "Made with Carrd" footer, Typedream's compulsory badge on its one-page free tier — is annoying but survivable; it tells a visitor the site cost nothing and that's it. A structural badge is worse: Webflow's Starter badge cannot be removed at any price on that plan, and Notion Sites ties badge removal to owning a custom domain specifically, so the "Built with Notion" mark sits on Free and on Plus alike unless you've also paid for the domain add-on.

The second is the export lock, and it's the one that costs you later rather than now. A free plan that lets you build something real but never lets you leave with it is optimized for retention, not for you. Butternut AI is the extreme case among everything we looked at: no code export, no WordPress export, no static export, on any plan, free or paid.

The third is the page or item count set below what a real result requires. Two pages on Webflow Starter, one page on Typedream's free tier — these aren't storage limits you might plausibly stay under, they're counts sized to show you the editor and nothing else. That's the tell: a storage cap you could live inside is a real constraint; a page count of one or two is a demo wearing a pricing tier's clothing.

Capacity gets renegotiated quietly, and it's worth checking rather than assuming

Free allowances aren't static, and the changes don't always arrive as an email. Lovable merged its build, Cloud and AI-feature credits into a single pool in June 2026 — previously separate buckets became one shared allowance, which changes how far a free account actually stretches without any single number on the pricing page moving. We didn't personally watch an allowance shrink mid-test on any of the fifteen, but that Lovable change is the right shape to watch for: the mechanic that quietly resizes a free tier is usually a recalculation of how usage is pooled, not an announced cut. Re-read the terms every few months rather than trusting the ones you signed up under — the same discipline the subscription audit piece argues for on paid tools applies here too.

Where free beats the first paid tier

Google Sites is the cleanest case: free is not a limited version of anything, because nothing paid is built for the person using it for free. Notion Sites makes an odder case by accident. Plus costs roughly $8 a month on the annual rate, and removing the badge needs the domain add-on on top of that — roughly $16 a month combined just to get a clean custom domain, against a free tier that already gives you unlimited sites and one working notion.site address. For someone who doesn't specifically need their own domain, the jump from free buys almost nothing. Carrd has a version of the same trap the other way round: Pro Lite is $9 a year, and it explicitly does not include a custom domain — Carrd's own documentation describes that tier as being "for users who don't use custom domains," which means the cheapest paid option is barely different from staying on Free. Prices checked August 2026.

What free costs you that isn't capacity

Support is the obvious one and the least interesting — nobody expects a support queue on a $0 plan. The two that matter more are retention and export. Squarespace's trial doesn't just expire; content past the 14 days is "marked for permanent deletion," which is a harder edge than a feature getting locked. And export is the whole story with Butternut AI: whatever you build lives on Butternut's hosting permanently, on every plan, because there is no route off it at any price. That's worth knowing before you build something you'd mind losing, on any tool — the full framework for weighing a free plan against what it might cost to leave later lives in what AI tools actually cost, and the credit-metered tiers specifically are unpacked in how credit-based pricing actually works.

The eight worth using at zero cost

Webstudio, for anyone who wants real control and doesn't need a custom domain yet — the free tier is close to the full product. Google Sites, for anything internal or informational, because paying was never the point of that product for an individual. Canva Websites, for someone who already thinks in Canva and wants five real sites rather than one. Wix, accepted on its own terms — permanent, no card, and honest that the banner is the price of not paying. Carrd, for a single link-in-bio or landing page where a custom domain genuinely doesn't matter yet. Framer, for testing whether its design-first editor suits you before a credit allowance becomes the constraint. WordPress.com's free tier, for anyone who wants to write now and move to self-hosted WordPress later without starting over, ads and all. And Notion Sites, for anyone already living in Notion who wants one page live today and hasn't decided whether a domain is worth the jump.

The pattern across all eight is the same: none of them ask you to imagine the paid version to get value out of the free one. Everything below this line asks you to imagine it, which is exactly the difference the three categories above are for.