In 2015, three French students had an unglamorous problem: every new client project seemed to demand the same backend chores. A site needed content fields. An app needed an API. Somebody had to wire up users and permissions. Pierre Burgy, Aurélien Georget and Jim Laurie were writing JavaScript for the front end, then reaching for conventional content systems when the job moved behind the screen. The tools worked, but their assumptions were built around websites. Their clients were beginning to ask for websites and mobile apps.
They made their own answer and put it on GitHub that October. The name was a small manifesto: Strapi is short for “Bootstrap your API.” The company followed in 2016. Today Strapi sells an open-source headless content management system: editors create structured content in an admin panel, developers expose it through REST or GraphQL APIs, and separate frontends decide how it looks on a site, in an app, or elsewhere.
- Free, self-hosted CMS at the center; paid editorial tools and enterprise controls around it.
- One content model can feed many channels, but each frontend must be built.
- Its decisive early move was choosing content management over general API building.
- Customers include Tesco, Moustache Bikes, Airbus, PostHog and Yuka.
A framework with an identity crisis
The first Strapi was part API framework, part CMS. That broadness sounds useful until it becomes a roadmap: which user should get the next feature? The company’s own account says its research split neatly in two. Roughly half the users were building custom APIs; the other half were managing content. Every product decision threatened to disappoint someone.
The founders tested a fake pricing page. Users showed more willingness to pay for content-management features. They also saw a wider shift toward headless CMS software. That combination moved Strapi toward the category it now occupies. The lesson is more practical than romantic: when a tool serves two jobs, find out which job creates a budget before expanding the feature list.
“We published a fake pricing page and realized users were more willing to pay for content management-oriented features.”Strapi company handbook
There had already been a costly rehearsal. Early users wanted to extend the software, and the initial architecture was not the answer. The team substantially rebuilt it for the v3 alpha in 2017, with plugins and a new admin interface. The “alpha” label then became its own obstacle: Strapi later said some users hesitated to put it into production until the beta arrived in 2019. Two years of feedback and a rewrite bought flexibility; they did not buy instant trust.
One newsroom, twenty-four doors
A headless CMS earns its keep when the same information has more than one destination. Tesco offers a concrete example. Its internal employee portal sends content through 24 web channels to an audience the case study puts at 500,000 users. About 80 people work in its Strapi admin panel. Rather than force every channel into one page template, Tesco and implementation partner VirtusLab built an extendable system on Strapi Enterprise.
They did real engineering around the product: virtual tenancy, custom approval flows, a WordPress migration plugin, and single sign-on tied to Active Directory roles. Tesco reports that release cycles became roughly 50% faster and that the project was delivered in three months. Those are the customer’s figures, not a promise that installing Strapi will halve anyone else’s calendar. The useful idea is narrower: centralize content, keep audience permissions explicit, and build the channel-specific presentation only once.
Moustache Bikes, a French electric-bicycle maker, tells a smaller but equally revealing story. It says replacing WordPress with Strapi made publishing 100 times faster. That figure belongs to its own workflow. Still, the shape of the problem is familiar: a content change should not have to travel through a developer and a deployment queue every time an editor wants to update a page.
The price of having no head
Strapi’s distinction is control. The Community edition is MIT-licensed and can run on a customer’s own infrastructure. Developers can customize its Node.js code, extend it with plugins, model content in components and dynamic zones, and deliver it through APIs. A team can pair it with React, Next.js or another frontend without accepting a vendor’s page-rendering system. That attracts teams with developers and particular requirements; it can burden a team that really wants a ready-made website.
The commercial model follows that split. Community is free. Growth sells tools including Live Preview, Releases and Content History. Enterprise adds controls such as review workflows, SSO and audit logs, with support. Strapi Cloud provides managed hosting. A Cloud subscription covers hosting, while paid CMS features require a separate CMS license. For a buyer comparing quotes, that distinction matters more than the word “free” on the open-source tier.
A useful test before adopting it: list the content types, channels and editorial approvals you need. If the same content must appear in several products, a structured content hub can save repeat work. If the job is one simple brochure site, building and maintaining a separate frontend may cost more time than it saves.
The company itself has paid for its experiments. Its handbook says a nine-month fundraising roadshow in 2018 left its bank balance below zero before a $1 million pre-seed closed. A $4 million seed followed in 2019, a $10 million Series A in 2020, and a $31 million Series B led by CRV in 2022. The public funding story says plenty about ambition; it says little about what an individual deployment should cost. That calculation belongs to the buyer’s editors, engineers and hosting bill.
The editor returns to the room
Strapi 5, released in 2024, cleaned up APIs and brought stronger TypeScript support, while adding content history and refinements to draft-and-publish workflows. Later work made the editorial side more visible: live preview, in-context editing, a redesigned media library, and AI assistance for content models and metadata. This is the irony of headless publishing. Remove the fixed frontend, and editors soon ask to see what their work will look like anyway.
In 2026, Strapi opened another door: an MCP server that lets AI agents perform content actions under scoped admin tokens. Plugin authors can add tools; agent actions can appear in audit logs. The practical question is the same one Tesco faced with human users: who is allowed to change what, and how will anyone know? Strapi’s newest features are less a break from its origin than another attempt to make an API useful to the people responsible for what flows through it.
The company began by removing repetitive work from developers. Its more durable proposition is that content should have a home independent of any single screen, and editors should be able to run that home without asking for a code change. The API was the beginning of the story. The newsroom is why it became a business.