Webflow to a Headless CMS: How to Choose Between Payload, Sanity, and Storyblok

The short version

A 50 to 100 page Webflow migration costs $5,000 to $25,000, and the destination CMS barely changes that number. Payload, Sanity, and Storyblok all accept your Collections, your references, and your rich text with roughly the same amount of work.

What the destination changes is everything after launch: the recurring bill, who owns the content, who is responsible when the database falls over at 2am, and how many people can log in without triggering the next pricing tier. Those are the questions worth arguing about, and they are not the ones the vendor marketing pages lead with.

This guide covers what each option actually is, what it actually costs as of July 2026, and, for each one, the case where it is the wrong choice. If you already know Payload is your destination, skip ahead to our Webflow to Payload CMS migration guide for the field-by-field playbook.

What "headless" actually changes

Webflow couples three things into one product: where content lives, how editors change it, and how the site renders. A headless CMS splits content away from rendering. The CMS becomes a content backend you query over an API, and you render the site yourself with a framework like Astro or Next.js.

That split is the whole point. You get real version control, a clean separation between content and presentation, control over how pages are built and cached, and a codebase an AI coding agent can read end to end. It is also what makes the destination choice consequential, because you are no longer picking a website builder. You are picking where your business content lives and what you pay to keep it there.

One correction worth making, because it shows up in a lot of migration advice: there is no Webflow "CMS item wall" forcing you out. Webflow's May 13 2026 change raised item limits rather than lowering them. The durable reasons to leave are per-seat pricing at $39 per full seat per month on yearly billing, the drop in included bandwidth from 100GB to 50GB for sites coming off the Business plan, and lock-in. Anchor on those. If someone sells you a migration on the basis of item caps, they have not read the changelog.

Payload

Payload is an open-source, TypeScript-first headless CMS. It is the closest match to Webflow's CMS mental model: typed Collections with fields, references between Collections, and rich text. The difference that matters is that Payload runs inside your own codebase. Schema, admin UI, content layer, and frontend can live in one repo you own.

Two things a buyer comparing vendors in 2026 needs to know, and both are frequently omitted:

Payload was acquired by Figma in June 2025. The software is MIT-licensed, so the code you have cannot be revoked, and self-hosting means a change of vendor direction does not take your site down. But it is a real input to a vendor risk assessment and you should weigh it rather than discover it later.

Payload Cloud has paused new deployments. There is no first-party Payload hosting product available to a new project as of July 2026. Any claim that Payload gives you "free hosting" is false. The license is free. Hosting is always a separate bill.

Payload also requires a database, and this is where "free" quietly stops being free. Free database tiers are not suitable for a client site: Neon's free tier suspends compute when it hits its limits, which in practice means the client opens the CMS to publish and it is down. Budget $8 to $30 per month for a managed database you can put in front of a client.

Add hosting on top: $0 to $20 per month on a static-first stack with Cloudflare Pages, or roughly $20 to $50 per month on Vercel with a managed database. If you are being paid to build the site, Vercel's Hobby tier is not available to you. Its terms define commercial usage to include a paid consultant writing the code and receiving payment to create or host the site. Anyone quoting you a Hobby-tier hosting cost for client work is quoting a plan they are not allowed to use.

Sanity

Sanity is a hosted, developer-first headless CMS. Content lives in Sanity's Content Lake and you query it with GROQ. Editors work in Sanity Studio, a customizable editing environment you configure in code. Real-time collaboration is genuinely good.

The pricing detail that gets overlooked: the free tier includes 20 seats, 10,000 documents, and 100GB of bandwidth, with Growth at $15 per seat per month above that. Twenty free seats is the most handoff-friendly free tier of the three by a wide margin. If your constraint is "the client has a marketing team and I do not want to hand them a bill on day one," Sanity solves that problem better than either alternative.

The trade-offs are real. Content sits on Sanity's infrastructure, not in a repo you own. Rich text uses Portable Text, so Webflow's HTML output needs a conversion step on import, which is a genuine line item in the project. And Studio is structured and form-based: it does not look like the page.

Storyblok

Storyblok is a hosted headless CMS built around a visual editor. Of the three it feels the most like Webflow to a non-technical editor, because editing is in-context and click-to-edit. Content is modeled as components and content types, which map cleanly from Webflow Collections.

Check the seat math before you commit. The free tier is one seat and 100GB of traffic. One seat works for a solo owner and does not survive a client handoff, because the developer and the client are already two people. The next tier is Growth at $99 per month for 5 seats. Treat $99 per month as the real price of Storyblok on a client project, not the free tier.

That is not a knock on the product. If keeping marketers fully self-sufficient in a visual editor is the top priority, Storyblok is the best answer in this comparison and $99 per month is a reasonable price for it. It just should not arrive as a surprise in month two.

Contentful (the enterprise fourth)

Contentful is the long-standing enterprise default: API-first, mature, well-supported. It is a capable destination and it belongs on the list for large organizations that already standardize on it.

We are not going to quote Contentful pricing here, because we have not verified current figures against their pricing page and a stale number in a vendor comparison is worse than no number. Treat Contentful as a procurement conversation rather than a checkout page. For most teams leaving Webflow, it is more platform than the problem requires.

Where each one is the wrong choice

This is the section most comparison posts skip, and it is the one that will save you money.

Payload is the wrong choice if nobody will own the infrastructure. Self-hosting means someone is responsible for uptime, backups, database upgrades, and security patches, forever. If there is no developer relationship after launch and no budget for one, you have swapped a predictable subscription for an unowned server, which is a worse position than the one you left. It is also the wrong choice if your risk model needs a vendor with a currently-selling hosting product and a support contract, given that Payload Cloud is not taking new deployments.

Sanity is the wrong choice if your editors need to see the page while they edit it. Studio is excellent at structured content and unremarkable at making a marketer feel at home. It is also the wrong choice if the Portable Text conversion is a poor fit for your content: if your Webflow rich text is full of custom embeds and unusual markup, that conversion is where the migration budget goes.

Storyblok is the wrong choice if you are cost-sensitive and small. The jump from one free seat to $99 per month is the steepest step in this comparison, and if you only ever need two editors you are paying for five. It is also the wrong choice if you want the content in your own repo, because it does not do that.

Contentful is the wrong choice for almost every site that is currently on a Webflow site plan. If you are not already an enterprise with a Contentful contract, you are unlikely to arrive at it by evaluating on the merits.

And headless is the wrong choice entirely if Webflow's canvas is genuinely how your team works and the per-seat bill is not actually hurting. Migrating is a real project with a real cost. "I have heard headless is faster" is not a business case. A rising seat bill, a sunset product you depend on, or a genuine need for a codebase you own is.

What it costs: the project versus the next five years

A typical 50 to 100 page Webflow migration runs $5,000 to $25,000 whether the destination is Payload, Sanity, or Storyblok. The work that costs money is identical: pulling the real data out of the Webflow Data API, modeling the schema, importing items while remapping references, re-hosting every asset off Webflow's CDN, rebuilding the frontend, and putting SEO-safe redirects in place before launch. Anyone quoting you a materially different price for different destinations is pricing something other than the work.

The recurring side is where the destinations diverge, and it is worth writing out honestly rather than implying Payload is free:

• Payload: no license fee, no seat fee. You pay for a database, realistically $8 to $30 per month for client-grade, plus hosting at $0 to $20 per month on Cloudflare Pages or roughly $20 to $50 per month on Vercel with a managed database.

• Sanity: $0 up to 20 seats, 10,000 documents, and 100GB bandwidth, then $15 per seat per month. Hosting for your frontend is still on top.

• Storyblok: $0 for one seat and 100GB traffic, then $99 per month for 5 seats. Hosting for your frontend is still on top.

Note what is common to all three: you are hosting the frontend either way. The honest $0 path there is Cloudflare Pages, which has unlimited bandwidth and no commercial-use restriction, with the caveat that Pages Functions count against the Workers free allowance at 100,000 requests per day. Vercel Pro at $20 per user per month is the floor if you want Vercel for a client site.

AI editing: the honest version

A lot of migration content, including some of ours previously, claimed that closed platforms cannot be edited by AI tools. That is no longer true and repeating it discredits the rest of the argument. Webflow ships an official MCP server, and Wix and WordPress have official servers too. An agent can read and write your Webflow content today.

The real difference is narrower and more defensible. Through a connector, an agent reaches your content through a partial, rate-limited, unversioned interface: it cannot upload images or assets to the media library, it cannot create or apply Interactions, visual inspection requires the Designer open with the bridge connected, and there is no git history, so there is no branch, no review, and no way to revert a bad bulk edit. Requests are capped per API key, so large content operations are slow by design.

In a repo, the agent gets the whole artifact. It can change a field type, migrate the content, update the template, run the build to prove the change works, and hand you a diff you can reject. That is the advantage, and it is an advantage Payload has over Sanity and Storyblok as much as over Webflow, because with Payload the schema and the frontend are the same codebase. If that verification loop is not something your team will use, this argument should carry less weight in your decision than the seat math above.

How to choose

Answer these in order and the decision usually makes itself.

1. Will anyone own the infrastructure after launch? If no, rule out Payload and pick a hosted option.

2. How many people need to log in? Two or more, and Storyblok's free tier is out. More than a handful, and Sanity's 20 free seats become the cheapest answer by a distance.

3. Do editors need to see the page while editing? If that is non-negotiable, Storyblok, and budget the $99 per month.

4. Do you need the content and the site in one repo you own and an agent can work on end to end? That is Payload, and only Payload of the three.

5. Is the Webflow bill actually a problem? If not, and the canvas is working for your team, staying put is a legitimate answer.

Our default recommendation is Payload, and the reason is specific: for teams leaving Webflow to get an owned, AI-editable codebase, it is the only one of the three that delivers exactly that. But it is a default, not a rule. If the answer to question one is no, or the answer to question three is yes, Payload is the wrong call and we will say so on the review.

Where to go next

If Payload is your pick, our Webflow to Payload CMS migration guide is the deep dive: field mapping, the step-by-step process, and what usually goes wrong. If you are still deciding whether to leave Webflow at all, start with the complete Webflow to code guide, then the cost breakdown so the budget is not a surprise.

Not sure which destination fits? That is what a free migration review is for. We will look at your actual Collections and your actual seat count and tell you which of the three fits, including when the answer is that you should stay where you are.