// the find
keithrozario/Klayers
Python Packages as AWS Lambda Layers
Klayers maintains pre-built AWS Lambda layers for popular Python packages, published to every AWS region on a monthly cadence. It solves the genuinely tedious problem of getting packages like pandas, requests, or spacy into a Lambda function without bundling them yourself. The target audience is anyone writing Python Lambdas who doesn't want to manage their own layer build pipeline.
Monthly automated rebuilds mean you're not stuck with a snapshot from 2022. ARM64 layers are first-class alongside x86, which matters for cost on Graviton functions. The Terraform provider and CDK integration mean you can pin to 'latest' programmatically without copying ARNs by hand. The history rewrite that dropped the repo from 200MB to 17MB was the right call — clone time on CI was genuinely painful before.
Layer expiry after 365 days is a real gotcha: if you hardcode an ARN and stop paying attention, your deploys will silently break when the layer gets deleted. The pipeline is Serverless Framework + Step Functions + DynamoDB — a significant amount of infrastructure to run yourself if you ever needed to fork or self-host. Package selection is community-driven via CSV pull requests, so if something obscure isn't listed you're on your own. No pinning to a specific package version — you get whatever was latest on the first of the month, which can introduce breaking changes if a package ships a bad minor release mid-cycle.