awslabs/project-lakechain

Provides 60+ composable middleware components for text extraction, transcription, summarization, and RAG pipelines, deployable as infrastructure-as-code using AWS CDK. Automatically scales document processing across Lambda, ECS, and SageMaker with GPU/CPU options, handling millions of documents while scaling to zero when idle. Supports custom middlewares for extensibility and includes 50+ ready-made examples covering video analysis, podcast generation, and multimodal AI workflows.

186 stars.