Skip to content

Latest commit

 

History

History
33 lines (22 loc) · 1.07 KB

README.md

File metadata and controls

33 lines (22 loc) · 1.07 KB

Amazon Bedrock architecture with Knowledge Base and Lambda using SST

This is a proof of concept that uses Amazon Bedrock to create a chatbot.

It uses:

  • SST to deploy everything to AWS.
  • Pinecone as vector database to use the free tier and avoid charges (you can use OpenSearch or a different vector database).
  • Lambda functions in Python via boto3 to communicate to Amazon Bedrock.

Architecture

Get started

Setup your IAM credentials: https://docs.sst.dev/advanced/iam-credentials

Create a .env in the root of the project with the Pinecone crendentials:

PINECONE_API_KEY=your_api_key
PINECONE_CONNECTION_STRING=your_host

Execute the following commands with pnpm:

pnpm install
pnpm bundle-prompt
pnpm bundle-sync-kb
pnpm sst deploy --stage dev