1.9 Million Downloads Per Week
Astro crossed 1.9 million weekly npm downloads in June 2026, representing 85% year-over-year growth. That rate makes Astro the fastest-growing framework by download velocity in the WebPulse dataset — not in absolute volume, where React and Next.js remain unchallenged, but in the rate at which new projects are choosing it. Vero Scale's June 2026 registry analysis places Astro's growth trajectory ahead of every other framework in the content-delivery category.
Download numbers are an imperfect proxy for adoption. CI pipelines inflate counts. Monorepos multiply them. But year-over-year growth rates normalize those distortions — the same CI pipelines and monorepos existed twelve months ago. An 85% increase in weekly downloads means substantially more projects are pulling Astro into their dependency trees than were doing so in June 2025. The trend line is unambiguous.
What Zero-JS-by-Default Actually Means
Astro's architectural proposition is simple and consequential: ship zero client-side JavaScript unless a component explicitly requires interactivity. Every other major framework — React, Vue, Angular, Svelte — begins with the assumption that the browser will execute JavaScript. Astro begins with the assumption that it will not. Interactive components load as isolated islands within a static HTML page, hydrated independently and only when visible.
Server Islands, introduced in Astro 4 and refined through 2025-2026, extend this model further. Dynamic content — personalization, authentication states, real-time data — renders on the server and streams into the static shell. The browser receives HTML. No client-side framework runtime. No hydration waterfall. No JavaScript bundle to parse, compile, and execute before the page becomes interactive. For content-heavy sites, the performance difference is structural, not incremental.
The security implications are equally structural. Zero client-side JavaScript means zero client-side attack surface. No XSS vectors in framework code. No supply-chain risk from client-side dependencies. No prototype pollution. The absence of JavaScript in the browser is not a feature — it is the elimination of an entire category of vulnerability.
Comparative Growth: Astro vs. Svelte
Svelte, the nearest comparable framework in the content-first category, reached 1.8 million weekly downloads in the same period — a 45% year-over-year increase. Svelte's growth is healthy by any standard. But Astro's 85% rate nearly doubles it. The two frameworks serve overlapping but distinct audiences: Svelte offers a compiler-first approach that produces lean JavaScript; Astro offers a content-first approach that produces no JavaScript at all unless instructed.
The divergence in growth rates suggests that the market is increasingly selecting for the zero-JS proposition over the lean-JS proposition. When the choice is between shipping less JavaScript and shipping none, organizations building content-driven properties — marketing sites, documentation, media, corporate presence — are choosing none. Astro's 60,000 GitHub stars and 2,909 annual commits indicate a project with the community velocity to sustain this trajectory.
The AI-Readiness Connection
AI agents parse HTML. They do not execute JavaScript. A page that requires client-side rendering to display its content is invisible to the majority of agentic crawlers operating in 2026. Astro's zero-JS default means every Astro page is fully readable by AI agents without rendering — the content exists in the initial HTML response. This is not a theoretical advantage. As agentic traffic approaches 57% of all web requests, the frameworks that serve complete HTML on first response have a structural advantage in discoverability.
Astro's 85% download growth and its 90 WebPulse score are not coincidental. The score reflects what the download numbers confirm: content-first, zero-JavaScript frameworks align with where the web is moving. Organizations selecting a framework for a new content property in 2026 are choosing the architecture that performs for both human browsers and AI agents simultaneously. The download data from Vero Scale suggests that this choice is being made at an accelerating rate.
