Skip to content

Releases: moebaca/personalized-aws-features

personalized-aws-features v0.1.0

Pre-release

Choose a tag to compare

@moebaca moebaca released this 26 Mar 12:41
3392b4a

personalized-aws-features v0.1.0-beta

Initial Beta Release

This is the first release of personalized-aws-features, a tool that cuts through AWS announcement noise by analyzing your actual AWS usage and delivering only relevant feature updates.

Features

  • Service Usage Detection: Automatically identifies AWS services you're using via Cost Explorer
  • AI (LLM) Filtering: Uses Amazon Bedrock to analyze and determine announcement relevance
  • Deduplication: Tracks previously seen announcements in DynamoDB
  • Flexible Deployment: Run as CLI tool or deploy as AWS Lambda function
  • Infrastructure as Code: Complete Terraform configuration for AWS deployment

Installation

See the README.

Known Limitations

  • LLM relevance matching is roughly 80-90% accurate
  • Only verified working with Amazon Nova Lite model (testing with other models like Claude ran into rate limiting through Bedrock)

Coming in Future Releases

  • OTEL Metrics
  • Improved LLM prompting for better relevance matching
  • Support for additional Bedrock models
  • More messaging integrations

Cost

Daily operational costs are approximately $0.05 with Amazon Bedrock.

License

MIT