Context
This is the sharpest differentiator. Live fork-to-many of a running microVM is something disk-only competitors (Cloudflare, Runloop, microsandbox) cannot do, and the one true peer (Morph) has zero public customer proof. The validated way to explain it is fork(2) for a whole computer.
Deliverable
- A page leading with
sb.fork(16) returning N live children, the ~27ms / ~3 MiB numbers embedded in the snippet.
- The fork(2) metaphor up top, then best-of-N, tree-of-thought, parallel test execution, A/B agent runs.
- An honest "fork vs snapshot vs rebuild" comparison (reuse the blog post framing).
Links
- Backed by mitos-run/mitos A1 (fork ergonomics) and A5 (density benchmark).
- Part of the use-cases hub (C1).
Done when
- Page is live with the runnable fork snippet and the comparison.
Context
This is the sharpest differentiator. Live fork-to-many of a running microVM is something disk-only competitors (Cloudflare, Runloop, microsandbox) cannot do, and the one true peer (Morph) has zero public customer proof. The validated way to explain it is fork(2) for a whole computer.
Deliverable
sb.fork(16)returning N live children, the ~27ms / ~3 MiB numbers embedded in the snippet.Links
Done when