I'm thinking of working on an open-source, serverless compute system designed specifically for Python applications. Think of it as a simplified, local-first alternative to AWS Lambda. It will offers the developer experience of tools like Celery or Dramatiq, but without the overhead of managing brokers, queues, or long-running workers.
The core idea is to create a tool that feels as simple as writing a `@task` decorator on your aysnc function like Celery, but with serverless principles: tasks scale automatically, idle resources are automatically shutdown, and everything can run locally during development. When you're ready for production, the same code can run in the cloud—without any rewrites.
Surprisingly, there isn’t a simple, Python-native system today that offers this kind of serverless compute experience locally, especially one that supports long-running tasks. I’m building this partly for my own learnings and productivity, but also exploring whether others might find it useful too.
Would love your thoughts. I have a url to put in your email as well if you want me to let you know when I get past the ideation phrase and post something on github.
The core idea is to create a tool that feels as simple as writing a `@task` decorator on your aysnc function like Celery, but with serverless principles: tasks scale automatically, idle resources are automatically shutdown, and everything can run locally during development. When you're ready for production, the same code can run in the cloud—without any rewrites.
Surprisingly, there isn’t a simple, Python-native system today that offers this kind of serverless compute experience locally, especially one that supports long-running tasks. I’m building this partly for my own learnings and productivity, but also exploring whether others might find it useful too.
Would love your thoughts. I have a url to put in your email as well if you want me to let you know when I get past the ideation phrase and post something on github.