Marble Review (2026): Pricing, Features & Honest Verdict

Marble is an open-source headless CMS built specifically for blogs, changelogs and articles. It is the cheapest credible option we have tested this year, and the free tier is good enough to run a real publication. Best for: developers who need a blog behind a Next.js or Astro site. Price: From $5/mo (free plan: yes). Rating: 7.5/10.

What is Marble?

Marble is a headless CMS that does one job. You write posts in a clean editor, upload images, and pull everything into your own frontend over a REST API. There is no page builder, no template system, no theme marketplace. The team describes it as “super simple headless CMS” and the product actually holds that line, which is rarer than it sounds in this category.

The licence is worth reading before you commit. Marble is AGPL-3.0, which is network copyleft: run a modified copy as a service and you owe your users the source. For the overwhelming majority of readers, who will use the hosted version to serve a blog, that is irrelevant. If you were planning to fork it into a product you sell, it is the first thing to check with a lawyer.

The project has been in public development since November 2024 and the GitHub org was created in March 2025. The repo sits at 382 stars and 55 forks, licensed AGPL-3.0, written in TypeScript, with the last push on 22 August 2026. Two people carry most of it: taqh with 788 commits and mezotv with 173, plus a handful of outside contributors. It hit Product Hunt on the 25 August board. The homepage lists OpenCut, Helix DB, Databuddy and Candle as users, and carries a testimonial from James Perkins, CEO of Unkey.

What are Marble’s key features?

A REST API that works with the framework you already use

Content comes out over a plain REST API, and the docs carry named integration guides for Next.js, Astro, TanStack Start and Framer. There is an official TypeScript SDK and an MCP server, so you can point a coding agent at your content without writing a client. This is the whole product in one sentence: your posts live somewhere sane, your site fetches them, nothing else is imposed on you.

The documentation is hosted on Mintlify and covers the awkward parts rather than just the happy path. There are recipes for reading time and syntax highlighting, pages on custom fields and the content model, and an import and export section, which matters more than it sounds. A CMS you cannot get your posts back out of is a hostage situation, and Marble publishes the exit before you ask for it.

Media hosting with a CDN, included in the plan

Images and video upload directly into Marble and serve from a global CDN. You can also link to your own storage if you would rather. The storage numbers are modest (1GB free, 5GB on Hobby, 10GB on Pro) but for a text-first blog with a header image per post that is a lot of posts before you notice.

Webhooks that fire on content changes

Every plan gets webhooks, and they are metered rather than gated: 100 events a month on Free, 500 on Hobby, 1,000 on Pro. That is enough to trigger a rebuild on publish, ping a Slack channel, or kick a cache purge. Most CMS products in this price band put webhooks behind the top tier or an enterprise call. Getting them on a zero-dollar plan is the difference between a toy and something you can wire into a real deploy pipeline.

Readability scoring on the paid tiers

Hobby and Pro add what Marble calls AI Readability insights, a real-time score plus optimisation tips as you write. It is the one feature that is not available on Free, and it is the least interesting thing here. If you are choosing Marble, you are choosing it for the API and the price, not the writing assistant.

How much does Marble cost?

Marble starts at $5 a month on the Hobby plan, with a real free tier below it and a $20 Pro tier above. Billing is per workspace, not per seat, which is where the value sits. Sanity’s Growth plan is $15 per seat per month, so a five-person team pays $75 there against $5 here for the same headcount. Hygraph’s Growth plan starts at $199 a month. Marble Pro is $20.

Two things about the ladder are worth checking before you pick a tier, and neither is advertised.

The annual toggle does nothing on Hobby. Flip the pricing page to Yearly and Hobby reads $60 a year. Twelve months at $5 is also $60. You pay the identical amount and hand over a year of cash up front for no discount. Pro is different: $200 a year against $240 monthly, a real $40 saving that works out at two months free. If you are on Hobby, stay monthly.

Pro costs four times Hobby and doubles every limit, not quadruples them. Storage goes 5GB to 10GB. API requests go 25,000 to 50,000. Webhook events go 500 to 1,000. Team members go 5 to 10. Divide it out and Hobby buys you 5,000 API requests per dollar while Pro buys 2,500. Every metered thing on Pro is worth exactly half per dollar what it is worth on Hobby. The only new capability Pro adds is unlimited authors instead of five.

The step from Free to Hobby is the opposite shape and it is the reason we rate the ladder well overall. Five dollars multiplies storage by five, API requests by five, webhook events by five, authors by five and team members by five. Nothing in software usually scales that cleanly, and it means the decision to start paying is easy while the decision to keep climbing is not.

So the honest read is that Hobby at $5 is the best-value plan on the page, Free is a legitimate place to run a small blog, and Pro is what you buy when you have more than five authors rather than because the quotas got cheap. Pro includes a 3-day free trial. Payments run through Polar, and refunds are available within 7 days by emailing support.

Who is Marble best for?

Use Marble if you are a developer who needs a blog or changelog behind a site you already control, and you have been putting it off because every CMS you looked at wanted $200 a month or three days of setup. Use it if your team is under ten people and per-seat billing has been quietly eating your budget. Use it if you want the option to self-host later, since the whole thing is AGPL on GitHub.

It also suits the specific case of a developer tool company that needs a changelog. That is a content type nobody builds a CMS for, it always ends up hand-rolled in MDX, and it always rots. Marble handles it with the same API as posts, and the free tier covers it outright.

Skip Marble if you need structured content beyond posts and media. There is a custom fields feature, but this is not a general content platform and it does not pretend to be. Skip it if you are running a large editorial operation with review workflows and granular permissions, because Sanity and Contentful have spent years on exactly that and Marble has not. Skip it if 10GB of media is a ceiling you would hit, because there is no tier above Pro to move to.

Best Marble alternatives

Sanity is the obvious step up. The free plan covers up to 20 user seats and two datasets, and Growth is $15 per seat per month. You get a far deeper content model, a real query language, and a much bigger ecosystem. You also get per-seat billing, which is the thing Marble is undercutting.

Hygraph has a Hobby plan at $0 always and jumps to $199 a month for Growth. It is federated, GraphQL-native, and aimed at teams unifying content from several sources. If your problem is one blog, the gap between $20 and $199 is hard to justify.

Payload is the closest philosophical match: open source, developer-first, code-defined schemas, self-hostable. It was acquired by Figma, which is either reassuring or worrying depending on your view. Payload gives you far more control over the data model. Marble gives you a blog working this afternoon.

Final verdict: is Marble worth it?

Yes, and the free tier means you can find out for nothing. Unlimited posts, 5,000 API requests a month and 1GB of media is a working blog, not a trial. If you outgrow it, $5 multiplies every one of those limits by five. For a category where the standard move is to charge per editor and start the real plans at three figures, per-workspace billing at $5 is the whole argument and it is a good one.

What impressed us most is the restraint. Every CMS in this space eventually grows a page builder, a visual editor, a personalisation engine and a sales team. Marble has ten public repos, a docs site, an SDK and a price list, and the product still fits on one screen. Whether that survives contact with growth is the open question, but right now it is the reason the thing is pleasant to use.

The reservations are about size and headroom rather than quality. Two contributors carry the codebase, the top paid plan caps at 10GB with nothing above it, the terms of service have not been touched since September 2025, and the yearly toggle on Hobby is a discount that is not a discount. AGPL also matters if you plan to embed Marble in something you ship. None of that should stop a developer who wants a cheap, clean, API-first blog. It should stop a company that needs a content platform to still be there, and still be growing, in five years.