I know it's way too early, but maybe as a sort of early idea and in case anyone knows anyone. For example, Cloudflare is extending their workers offering to other languages like Python and Rust. Given fast startup and build times roc might lend itself towards serverless in future as well. Considering they have a pretty holistic ecosystem and Roc's platform concept, would it make sense to approach them as Roc approaches 1.0 to create a possible Roc serverless environment? Could make Roc accessible to more users and specific use cases.
If we were close to 1.0 I'd agree, but we're almost a year away from our first version ever, 0.1
So I don't think we should approach them until we our confident ourselves in our stability
Which I think might take a couple years
I expect the serverless ecosystem to revolve around WASM since absolute performance generally isn't a priority and the security benefits are significant. I think Roc is positioned well to produce small/fast WASM binaries (I guess that's the name?) and should be a good fit. Assuming WASM modules are stabilized and they provide a WIT for their environment I don't think there's anything that would require direct coordination.
Oh yes, if you're talking about Roc generating WASM, then we're already a great fit
I setup a runtime to run Roc on AWS Lambda using Basic CLI in this repo https://github.com/isaacvando/roc-aws-lambda-runtime
It’s very out of date so I’ll try to update it later today
Luke also wrote a Zig platform for running on Cloudflare (also out of date) https://github.com/lukewilliamboswell/roc-serverless
Isaac Van Doren said:
I updated the custom Lambda runtime to basic-cli 18 and the slowness I had seen in the past is now gone :smiley:. I'm very pleasantly surprised that AWS supports writing Lambda functions in Roc in such a lightweight way.
New repo url for the runtime: https://github.com/isaacvando/roc-aws-lambda
Last updated: Jun 16 2026 at 16:19 UTC