Automattic's Spacefast hosts vibe-coded apps for $5 a month

Share
Automattic's Spacefast hosts vibe-coded apps for $5 a month

Automattic launched a new company today called Spacefast, and the interesting thing about it is everything it refuses to do. It does not generate code, it does not offer a chat box, and it does not care which model you used. It takes whatever your agent already made — a page, a deck, a small app — and puts it on the internet at a private address, with a database, version history and a delete button you can walk back. Free to start with a dollar of monthly usage credit, five dollars a month for unlimited spaces, fifteen for a team tier.

That is a narrower product than the pitch decks in this category usually describe, and it is aimed at the exact spot where vibe coding stops being fun: the part after the code works on your laptop.

The half of vibe coding nobody priced

Generation is a solved-enough market with a revenue line attached. Lovable passed a $600 million run rate in September at a $13.3 billion valuation, with two-thirds of the Fortune 500 inside its customer list, as we covered when Lovable's run rate passed $600 million. Whatever you think of the multiple, the demand is not hypothetical: people who cannot read code are shipping things other people use.

The gap is the last mile. A deployed app needs a host, a database, a login system, a way to send a password reset, and someone to decide who is allowed to see it. Vercel — the default the coding agents reach for — assumes a repository, a build command and a person who understands what an environment variable is. Batuhan İçöz, who runs Spacefast, told TechCrunch that the idea came from non-technical friends sending him Vercel links while they explored vibe coding, and from a simpler question: Automattic already owns a CDN, so why route through someone else's? He built the first version as a side project called Stattic during an internal Automattic sprint, and the company liked it enough to make it a separate business rather than a WordPress feature.

That last detail is the tell. Spacefast is not a product line inside WordPress.com. Automattic spun it out as its own company, per Automattic's launch announcement, and it runs on WordPress.com's infrastructure — the platform Automattic says serves more than 15 billion page views a month.

What actually shipped, and what didn't

The mechanics are more conventional than the positioning. You can publish from a chat with a connected agent, by dragging a folder onto a web page, by pushing to a Git branch, or from a command line and HTTP endpoint. Supported stacks are Vite, TanStack Start, Next.js, Astro and PHP; anything static works, and app runtimes depend on the path you use. Every publish becomes a version with its own permanent address, and rolling back is a pointer flip rather than a rebuild. Spaces are private by default, and comments attach to elements on the page so a client can point at the chart they hate and your agent can go fix it.

The distribution is the part worth studying. Spacefast ships an MCP server and appears as an official plugin in the Claude directory, the ChatGPT plugin store, the Cursor marketplace and GitHub's MCP registry; the setup page lists about nineteen agents, Hermes included. That is the same plumbing described in AI 101 — What is MCP?, used as a sales channel instead of an integration chore. Spacefast never has to win a user away from their chatbot. It has to be the thing the chatbot offers when the user says "put this online."

The honest gap is what is not live yet. Automattic's release says built-in app authentication, transactional email and shared team knowledge — brand context, playbooks, the stuff an agent should read before it starts — are "launching in the coming weeks." TechCrunch's writeup lists authentication among the features Spacefast provides. Today you can publish and share; a real product with logins and password resets is still a few weeks out, which is the difference between a demo host and a place to run a business.

The bet: own the rails, not the editor

Spacefast's pricing tells you who it thinks the customer is. The free tier includes twenty spaces, custom domains and a database. Unclaimed spaces stop thirty-three hours and twenty minutes after the last publish, capped at thirty days from creation, according to the product's own documentation — a rule that makes anonymous publishing frictionless and makes keeping anything a paid act. The five-dollar tier removes the space limit. The fifteen-dollar tier adds members at five dollars each after ten.

Charging for infrastructure rather than tokens is the strategic choice. Lovable, Emergent and the rest bundle the builder and the rails and bill you for generation; Spacefast lets you keep paying OpenAI or Anthropic directly and charges only for the running. That makes it a Switzerland play, and Switzerland plays need two things: enough integration surface that no agent is a second-class citizen, and enough enterprise trust that procurement stops asking questions. Automattic has the second one. Spacefast's enterprise pitch is sprawl — marketing teams quietly building landing pages on Lovable until IT notices — and the counter is that WordPress VIP and Pressable customers can adopt it without onboarding a new vendor.

There is a cannibalization story here too, and Automattic is telling it on purpose. Matt Mullenweg called the project "the future of WordPress and AI" when it soft-launched in September, as wp-content.co reported, and İçöz told TechCrunch that some sites simply will not need a content management system. The company that sold the world a CMS is now selling the layer underneath sites that never had one.

Who loses, and the skeptic's case

The obvious losers are the low end of hosting: the shared-hosting resellers, the hobby tier of Vercel and Netlify, and every "deploy your AI app" startup that appeared this year. Cloudflare open-sourced a vibe-coding platform for non-coders in August, per Ars Technica, and the big platforms all have agent-facing deploy paths now. Publishing an artifact is becoming a checkbox feature, not a company.

The skeptic's case against Spacefast specifically has three parts. First, churn: İçöz says out loud that he expects people to leave, because most of them were never that interested in making apps. A platform whose free tier expires in under a day and a half is optimized for trials, not retention, and the number that would settle it — what share of spaces get claimed — is undisclosed. Second, switching costs are near zero at the low end: the product's own FAQ notes that static sites usually move to another host as they are. Third, timing: Automattic spent the autumn in boardroom chaos, with Mullenweg pushed out and back within days, an interim finance chief gone weeks later, and layoffs across its properties in October. Spinning out a new company in that window can be read two ways, and only one of them is about focus.

What to watch

Whether authentication and transactional email arrive on the promised timeline is the first test, because those are the features that separate hosting from a platform, and they are also the ones that break. The second is placement: the Claude and ChatGPT directories are ranked lists, and being the default publish target in a chat window is worth more than any feature comparison. The third is what Automattic does with its own channel — if WordPress VIP and Pressable start bundling Spacefast, the enterprise thesis is real; if the agency workshops stay the only bridge, it is a side business with good branding.

And watch the incumbents. If Vercel ships a one-click "publish this artifact" flow aimed at non-developers, Spacefast's advantage narrows to the CDN, the procurement relationships and the fact that Automattic will answer the phone. That is a real moat. It is not a $13 billion one.

Would you trust a host that deletes your project after 33 hours unless you claim it? Write to us via our contact page.

Get SiliconNoon in your inbox

One email a day with the AI stories that matter — what happened, the numbers, and why it matters. No spam, unsubscribe in one click.

Read more

Deep Dive — CICC's 100 billion tokens a day, and no ROI number

Deep Dive — CICC's 100 billion tokens a day, and no ROI number

China International Capital Corporation burns between 70 billion and 100 billion tokens a day, and expects to hit 200 billion next year. The figure comes from the bank's own IT chief, Xie Bisong, at a Tencent Cloud briefing on October 28, and 21st Century Business Herald reported it under a headline that names the real story: the AI race in Chinese finance has moved from counting tokens to counting return. That return does not yet exist as a number.

Salesforce buys Braze for $5.66B at five times forward revenue

Salesforce buys Braze for $5.66B at five times forward revenue

Salesforce said on Tuesday it will buy Braze, the customer engagement platform, for about $5.66 billion in enterprise value, paying $47.50 a share in cash — roughly 58% above Braze's $30.04 close on Monday, per the announcement. Braze stock jumped more than 50% and was briefly halted premarket; Salesforce drifted about 1% lower, CNBC reported. The deal is expected to close in Salesforce's fiscal 2028, which starts in February. The interesting number is not the price.

Microsoft named a segment "Agents and Infra." Read the fine print

Microsoft named a segment "Agents and Infra." Read the fine print

Microsoft did two things on Wednesday that look like bookkeeping and are not. It told investors what Azure earns — $29.42 billion in the June quarter, a figure it had refused to publish for a decade — and it collapsed three reporting segments into two, one of which it named "Agents and Infra." Then it quietly took away the segment-level cost and margin detail it had been publishing all along. You now know the top line.

Deep Dive — Anthropic says recursive self-improvement may be near

Deep Dive — Anthropic says recursive self-improvement may be near

The most important sentence in Anthropic's amended IPO filing is not a number. It is a forecast, buried in the risk factors, that the company's own next models may cross the line it drew for AI that can meaningfully speed up its own development. "Our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Policy's definition of 'dramatically accelerated progress attributable to AI,'" the…