The best free website builders

Carrd answers a question most of this category has stopped asking: what if the site is one page and that is fine. A very large number of small businesses need an address, a phone number, a description and a photograph to exist online, and buying a platform with a content system and a store to achieve that is a mismatch of tool to task.

The economics follow from the restraint. Because there is one page, no store and no content model to run, the paid tier costs a few dollars a year rather than a month, and the free tier builds several sites without paying at all. Setting one up is a single sitting, and there is nothing to administer afterwards.

The cost is everything the restraint excludes. Selling, publishing, bookings and a second page are all absent, and support is a contact form with no stated response time. A business that grows into needing any of those is not upgrading within Carrd, it is leaving. For the reader whose site is genuinely one page, that is not a real objection; for anyone else it is the whole story.

Google Sites publishes a simple informational site for no money. A custom domain is documented in the consumer help with no plan named, which makes it one of very few products anywhere that puts a site on an address the reader owns at zero cost. For a club, a school page, an internal directory or a business that genuinely needs three pages and a phone number, that is a real answer.

The editor is small and quick, with a theme covering images, fonts, buttons and menu layouts.

What you give up is everything beyond that, and it is missing outright rather than gated. There is no store, no blog, and no search-engine controls at all — the single search setting asks engines not to display the site. Support is documentation only for a free account. And the picture is muddied by the vendor's own material, which elsewhere sells its paid subscription as the route to "more control, like linking to your custom domain", so the free route is documented and simultaneously undersold by the company that documents it.

Blogger is the rarest thing in this category: a product that publishes on a domain you own, with a certificate the platform creates and renews, and charges nothing for any of it. There is no rate card, no upgrade path and no plan that unlocks the thing you actually wanted. What it does, it does for everyone.

Underneath, it is a blog and only a blog. Posts and photos are stored and served by Google, the design surface is a template picker with editable layouts and background images, and revenue is meant to come from advertising rather than from you. Everything can be taken out again as a feed file, and the theme downloads separately, so nothing you write is trapped.

What you give up is everything that has happened to publishing since. There is no store, no newsletter, no audience tooling beyond visit statistics, and no documented way to reach anyone at the company — Blogger scores 1 for support because the help centre's only invitation is to leave feedback on an article. Connecting a domain means entering two CNAME records and four A-records by hand at your registrar. The product is maintained rather than developed: the blog has not posted since 2020, while the Android app and the help centre are both still being updated.

Direct.me is priced to be an easy yes. There is one paid membership rather than a ladder, it unlocks every block, element, app and AI action the product has, and paying for a year costs about what two months of the monthly rate would. Around it sits a free plan that publishes a working page, with individual blocks marked as either included or carrying an upgrade prompt.

Setup is meant to take minutes and mostly does. The onboarding is three steps — claim a handle, add content, share the link — and the vendor ships whole vertical presets it calls collections, described as installed and ready in about a minute. What the product is sold on beyond that is owned audience: email and contact capture on the page itself, live analytics, lead tracking, and calendar and customer-record integrations once you pay.

The thing to weigh is how little of this can be checked. The vendor's documentation and pricing set sits behind a challenge page that will not open, so the design ceiling is described in marketing language and nothing bounds it in either direction. Direct.me scores 2 for design and 2 for support on that basis, both numbers a reader should treat as provisional rather than settled.

Hashnode is a blogging platform with a distribution argument attached. Posts go out on your publication and also into the platform's own network feed, which is the part a new technical blog cannot manufacture on its own. Above that, the paid tier adds machinery aimed at how software writing gets found now: structured question-and-answer data, a machine-readable index file, a staleness queue and daily tracking of whether the major assistants are citing you.

The writing surface is the whole product: a technical editor or raw Markdown, with an import that takes a zip of Markdown files and maps tags, canonical addresses and publish dates across, so an existing blog arrives intact. Search settings are careful for a hosted platform — the title is stored separately from the blog name and crawler access is switchable.

The catch is that nothing establishes how far the appearance can be changed. No page read documents a theme editor, a template count or a styling surface, so design is scored at 2 on the evidence available rather than on a ceiling anyone found — and no support channel, response time or help centre appeared on any page either. What is documented instead is a fast public changelog, which says the product is alive without saying who answers the phone.

Systeme.io does the thing this category usually charges for and gives it away. The free plan carries a domain you own, funnels, a blog, a store with inventory and shipping, a booking calendar, courses and a list of up to two thousand contacts, and the vendor states it never expires and needs no card.

The branding question, which is what usually spoils a free plan, resolves cleanly. The mark on a page is an opt-in affiliate badge you switch on to earn commission rather than a watermark you pay to remove, so a free page can be entirely yours. A site can be published at the root of your domain, and there is a free setup service linked from every page for anyone who does not want to touch DNS at all.

The tradeoff is that nothing leaves. No export is documented anywhere — not the pages, not the funnels, not the courses — which holds growth at 3 despite a feature list that would otherwise reach higher. The editor is a plain drag-and-drop tool over a template library, with no AI anywhere and no documented site-wide styling surface, so design scores 2. The pricing FAQ also contradicts the rate card on the same page.

Canva Websites exists because a very large number of small business owners already open Canva to make a flyer, and asking them to learn a second tool to make a web page is a worse answer than adding a publish button to the first one. The page is laid out exactly like any other Canva design, and publishing it costs nothing on a Canva address.

What makes it work is the absence of a second mental model. Elements are dragged, resized and aligned with the same controls used for print and social work, the brand kit carries colours and fonts across, and there is no separate concept of templates, sections or a content system to understand before a page exists.

It is not a website platform in the sense the rest of this category means. Selling, publishing and any structured content model are all absent, so a business whose site needs to do more than describe the business will outgrow it quickly. Search settings are better than the rest of that list suggests — per-page titles and descriptions, editable addresses and an automatic sitemap are all documented. As a fast, good-looking single page for a reader already inside the design tool, it is hard to beat; as the foundation for a growing site it is the wrong starting point.

Framer occupies the space between a builder that constrains and a tool that overwhelms. The canvas is free — elements go where they are put — but the defaults are good enough that a page looks current before much has been decided, and motion and interaction are configured rather than coded.

The speed comes from that pairing of a free canvas with strong defaults, plus a content system for anything repeating. Its entry plan includes a custom domain rather than treating that as an upgrade, which removes the most common surprise in this category, and the free tier is a genuine place to build rather than a countdown.

The trade-off is scope. There is no store, no bookings and no marketing email, so a business that sells or schedules will be assembling those from other services and linking out. Framer is at its best on the marketing side of a business — a launch page, a portfolio, a site whose job is to look sharp and explain — and noticeably less suited to being the system a business is run from.

Umso positions its editor as deliberately limited: fewer settings, fewer choices, and a shape suited to a small marketing site rather than to anything that might grow in an unexpected direction. For a founder who wants a presentable site this afternoon and does not want to make thirty design decisions to get one, the narrowness is the feature.

It is not a one-page tool, despite what the free tier suggests. Only the free plan is limited to a homepage; the first paid tier publishes up to ten pages and the tier above removes the limit entirely.

Pricing is per site rather than per account, which is cheap for one project and adds up quickly for anyone running several. Documentation is the other difficulty: the pricing page, the features page and the help centre all returned too little to read, so several capabilities are recorded as absent from the homepage's own feature list rather than as confirmed absences. That is a thinner basis than a published statement would be.

Substack collapses a blog and a mailing list into one object. A post is an email and a web page at the same time, subscribers are the audience and the list, and charging for some of that writing takes no extra setup. For a consultant or writer whose website is mostly words plus an email list, the assembly that usually takes three products takes none.

It is a real site rather than only an archive: an about page and other custom pages sit in the navigation, and a domain the writer owns can be attached for a single charge rather than a subscription.

The cost comes in two parts, and both are worth stating plainly. The platform takes a tenth of every paid subscription, permanently, on top of the payment processor's fee — which is nothing while nobody is paying and becomes the largest line item once they are. And design control is close to absent: a publication looks like the platform, with a handful of choices around a fixed shape, which is the price of the writing being the product.

Bio.link is shaped around a decision the vendor made once: there is a free plan and there is one paid plan, and the paid plan carries everything. Unlimited pages and visitors, a domain you own, removal of the vendor's mark, the email list and the posts all arrive together at $7.49 a month on yearly billing. Nothing is discovered two tiers up.

Getting there is the other half of it. Setup is a theme, a stack of links and a publish, described by the vendor in seconds and by the people quoted on its own pages in a day or so. The page can be updated from a phone, the design controls are themes with editable fonts and colours, and an AI assistant will draft and answer on your behalf once you have written it a knowledge base to work from.

The cost of that simplicity is a ceiling you meet quickly if you want the page to be more than a page. There is no per-block styling, no stylesheet and no custom HTML anywhere in the product, so Bio.link scores 2 for design. Search visibility is asserted as a property of the output rather than offered as controls you edit, and the free plan carries the vendor's branding and no domain of your own.

Campsite's free plan is the argument for it. Unlimited links, the ability to build a theme from scratch rather than pick one, and analytics on what people click all arrive without paying, which is not how this category usually works. The upgrade at $7 a month adds a domain you own, removes the vendor's mark, and unlocks the premium link types and the marketing integrations.

The link types are where the product spends its detail. There are around twenty of them, thumbnails can come from an upload or a stock library, and the icon set runs to four figures. Behind the page, reporting covers views, clicks, click-through rate, location, referrer and device, with history rising from a fortnight on the free plan to lifetime higher up.

The cost is that everything above the surface is fixed. There is no typography control, no stylesheet and no code block anywhere in the vendor's material, which puts design at 2. Extra profiles and collaborators are metered at $1.50 each per month, so a person running several pages pays more than the headline suggests, and the zero percent taken on tips is described as being for a limited time rather than as a standing arrangement.

Fourthwall removes the parts of selling physical products that stop most people starting. There are no product minimums, no coding, and no manufacturing to arrange: you design the item, the platform makes it, ships it, answers the buyer and registers and remits the sales tax. Its revenue comes out of what sells rather than from a subscription, so a shop that sells nothing costs nothing.

The free plan is the real plan. A domain you own connects on it, the shop can be designed and published, and the deduction depends on what you sell — a listed flat cost on catalogue products, nothing on items you ship yourself, a percentage on digital goods and memberships. Beside the shop sit memberships with members-only video, a custom mobile app, promotions, coupons, donations with no added fee, and listings pushed onto the big video and social platforms.

The cost is that the design surface stops at a theme. You choose one, preview it live, arrange sections per page and set styles inside them; there is no free canvas and no stylesheet, so design scores 3. The percentage on digital products and memberships is only removed by moving to the paid plan, which also adds the round-the-clock priority channel — on the free plan the platform answers your buyers rather than you.

Ko-fi is built around a single decision: it never holds your money. Payments go straight into your own processor account, which means there is nothing to withdraw, no payout fee and no delay. Its own cut is capped at five percent and is nothing at all on one-off tips for a creator who never subscribes, with the processor's charge stated openly beside it.

There is no build step. You create an account, connect the payment account you already have, and share the link. What sits behind it is wider than the setup suggests: a catalogue shop for digital and physical items with shipping calculation, order tracking and discount codes, memberships, a post feed with media, pinning, polls and scheduling, stream overlays and chat bots, and embeddable buttons for a site you own already.

The cost is that this is a payment page rather than a website. It publishes only inside the vendor's own address — there is no domain of your own on any plan — and the appearance controls are a cover image, one colour, a social preview and the word used for a tip, which puts design at 1. Support is a ticket form with no published response time and no live channel, and there is no way to edit search settings anywhere in the product.

Komi looks like a link page and is sold as a business development tool. Behind the mini-site sits a marketplace that matches creators to brand deals, a customer record holding the audience you capture, an advertising pixel, performance reporting and a paid-reach product. That is what the product is actually selling, rather than the page itself.

Setting the page up takes minutes: claim a username, choose a template, stack the modules you want and share the link, all of which works on a phone. Selling happens on the page — digital products, custom products, tipping and a connected shop all check out without sending anyone elsewhere. Support is an in-house team the vendor says a creator always reaches a person at.

The cost is the page. The marketing calls the mini-site fully customisable and the support centre's own instructions bound that at a header image, a logo or display name, brand colours and an overlay toggle — no typography, no stylesheet, no code block and no layout control anywhere, which puts design at 1. The templates on offer are content starters rather than design themes. A free plan publishes a complete page, and the domain, the branding removal, the pixel and all email sending sit behind the one paid tier.

Lemon Squeezy takes on the part of selling digital goods that has nothing to do with the product: tax. As merchant of record it registers, collects and remits sales tax in the places you sell, which is the obligation that quietly turns a side project into an accounting problem. Its cut covers that rather than only covering payment processing.

There is no monthly charge and no plan ladder, so a store that sells nothing costs nothing, and a domain you own attaches without any tier to reach. The commercial tooling around it is deeper than the size of the product suggests: email broadcasts, discount and coupon codes, affiliates, lead magnets, pay-what-you-want pricing, recovery for abandoned carts and failed payments, and revenue reporting.

The cost is that this is a settings screen rather than a website. Design control is a store logo, a header image and product imagery — no page editor, no template gallery and no stylesheet, which makes it a 1 for design, the narrowest surface among the stores here. The percentage is unavoidable because it is the business model, and it rises for extra payment methods and cross-border sales. Support is promised as staffed but no channel, tier or response time is published anywhere, and support takes a 2 for it.

Milkshake is designed for the person whose whole business runs from a phone. There is no browser editor and no desktop version — you build, publish and update from the app, which is why the usual hosting and DNS obstacles never appear. A permanent free plan publishes a complete site with every card type and every Look available, and the vendor states plainly that it is not a trial.

Customisation is a closed set applied to whichever Look you pick: up to five colours, a font per text section with weight and case options, and a background pattern for the desktop rendering. That is enough to look deliberate and not enough to look bespoke, so design scores 2. Selling digital files works on every plan including the free one, and the platform's percentage falls as the tier rises.

The catch is the ceiling on both axes. Growth is vertical rather than structural — more cards at one address, a handful of separate sites per account, and no team seats offered at any price on any tier. Search settings and analytics history are sold up the tiers rather than included. The domain sits only on the dearest of the four plans, billing runs through the app stores, and the vendor warns the published figures vary by country and platform.

Pensight is a storefront for selling time and knowledge, and the part it takes most seriously is the calendar. Availability is set weekly with configurable slot lengths, both major mail platforms connect, reminders go out by email and text, video calling is attached and calls are transcribed automatically. That is a scheduling product rather than a booking link on a page.

Around it sits a wide commercial layer: eight offer types through one checkout with coupons, instalments, pay-what-you-want and subscriptions, plus email broadcasts, upsell funnels, an affiliate programme with commissions you set, and the three major advertising pixels. Setup is a guided three-step flow with templates for each offer type, and the unpaid plan is real — no monthly charge, a percentage at checkout instead.

The catch is how that percentage is described. The headline says a single three percent transaction fee inclusive of card processing; the comparison grid on the same page splits it into a platform row reading zero and a processing row reading three. Both come to the same number for you, and a reader told zero would be misled. Beyond that, the profile page's controls are an enumerated closed list with no stylesheet anywhere, and a domain you own needs the Pro plan.

Setmore is a scheduling product first, and the page it publishes is a front door to the calendar. Setup is a form with a completion meter and a live preview beside it, with no domain, hosting or template decision anywhere — the page appears inside the vendor's own site as soon as the form is filled in. Staffed onboarding demos are offered to everyone at signup.

That support arrangement is the standout. Round-the-clock chat, email and phone are printed on the free plan card as well as the paid one, with phone numbers published for two countries, and it earns a 5 for support at a price where that is close to unheard of. The free plan itself is a standing arrangement rather than a trial: four users, a monthly appointment cap, no card at signup.

The cost is that the page is barely a page. Customisation is a banner, a logo, one brand colour, three button shapes, three theme modes and a photo gallery, with no template choice, font control or stylesheet anywhere, which puts design at 1. No plan connects a domain you own. And the per-user price counts every staff calendar, including ones nobody logs into, so a small team pays several times the headline figure.

Taplink is what happens when a link page keeps growing. There are three hundred ready-made templates, you can build a theme rather than pick one, every block can be styled individually, backgrounds take your own colours and images, the fonts are yours, and from the middle tier there is a block that takes raw HTML. That puts design at 3.

The commercial layer grows the same way. Page-view statistics are on the free plan, click analytics and the advertising pixels arrive with the middle tier, and the top tier adds a customer record, lead capture forms, automated emails, lead export, promotion codes and webhooks. The vendor's claim is a page built in minutes from a phone or a computer, and nothing in the evidence contradicts it.

The cost is that the setup burden climbs with the ambition. A page of links takes minutes; a page with forms, payments and a customer record behind it does not. Selling and the domain both sit on the top tier rather than lower down. And the price has a shape worth reading carefully: $8 a month is the twelve-month rate paid up front, the minimum commitment is three months, and there is no monthly option at all.