How Wasmer used Codex to build a Node.js runtime for the edge
Teams use Codex with GPT‑5.5 to take on more ambitious projects and complete them 10x to 20x faster.
10x to 20x
Increase in development speed
Engineers at Wasmer experienced a breakthrough this year: They figured out how to run Node.js workloads inside a WebAssembly sandbox, enabling developers to run JavaScript apps, MCPs, and agents without Docker. This effort would have taken one year without Codex, but with Codex, it took two weeks. Now, they’re the first cloud host to provide full Node.js at the edge layer.
“We are actually moving out of the IDE itself. We’re not touching as much the code, we are just guiding it where we want it to go.”
Wasmer is a young company with a small team and a big goal: an edge computing platform that scales across local and global environments without platform restrictions.
“The kinds of projects that we are taking on are incredibly ambitious, and it would have taken us ages to do them without AI,” says Syrus Akbary Nieto, Founder and CEO. “Now, things are way simpler and faster.”
That simplicity and speed translated directly into a groundbreaking project: Edge.js(opens in a new window), a JavaScript runtime that can run Node.js workloads for AI and edge computing.
It’s a project the team had long wanted to take on, but they didn’t always have the resources to do it. “Everyone here is very, very technical, but we just didn't have the time to dedicate to these projects. And now, we can launch new products that would have been impossible before,” explains Nieto.
With Codex, Wasmer engineers can multiply their efforts. “We have increased development speed by 10x to 20x, at least,” says Nieto.
“We were able to create a JavaScript runtime in just two weeks. Without AI and without Codex, it would have taken us easily one year.”
Like many teams, Wasmer engineers were initially skeptical of AI. “We were not very trusting of AI outputs at the beginning,” explains Nieto. But when the team started experimenting, the results exceeded their expectations. “Over the last year, and especially over the last few months, we have been working with Codex, and the results have been really, really good.”
As Codex’s reasoning abilities evolve, Wasmer engineers spend less and less time with handholding. “We are actually moving out of the IDE itself, and we are not touching the code as much. We are just guiding it where we want it to go,” says Nieto.
The team used Codex from the very beginning of the project to the very end, from building the initial architectural building blocks to polishing the final product. Throughout, Codex helped the team find bugs and identify their root causes.
“There were certain bugs that we didn’t imagine we were going to have, and as soon as we started discovering them, Codex went directly into debugging,” Nieto says. “The impressive thing for us was seeing how fast it went from debugging to finding the root cause and identifying the solution.”
Typically, Nieto explains, developers would need special expertise to find many of these bugs. But Codex, Nieto says, “was able to master console logs to trace calls and a low-level debugger like LLD, which accesses things on the assembly level. Codex can get very low level, and see what is happening under the code.”
“There are certain subtleties that we don't know of because we are not experts in C++. Codex was able to spot them pretty early.”
Before Edge.js, Nieto explains, it wasn’t possible to run a JavaScript runtime at the edge. And before Codex, it would have been impossible for the Wasmer team to take on a project this ambitious and with this timeline.
“Codex enabled a small company to achieve things that were only possible at big companies. This project literally would have been impossible without it,” says Nieto.
With their ambitions emboldened, the Wasmer team is looking ahead to even bigger projects. “Now, we have at hand things that were not possible before. We need to look at even more challenging problems,” says Nieto.
