Guide to Creating an AWS Bedrock Knowledge Base application
Step1 : Creating an IAM User:
1.????? Click on IAM from AWS console or you can type IAM in the search bar
2.????? Access Management Dashboard>User Groups>Create Group
3.????? Attach these policies to the User group – Otherwise you won’t be able to access data from the user or use OpenSearch knowledge base
4.Users>Create User
?5.Add user to group
6. user>Security Credential>Create Access Key> Other>Next>Create Access Key>Download CSV file and save credentials for later
?
?
Step 2 : Requesting Access to Bedrock FM models
1. Console>Amazon Bedrock>Model Access
?
2.Manage Model Access>Anthropic>Submit use case details > Request Model Access> Select all
?
?
Step 3 : Setting up AWS credentials in Github Codespaces using Boto3 (Setting up Environment Variable)
1. Settings menu>Code Spaces> Codespace user secrets> New secret> Add your AWS credentials here> Attach to the Repository > Add Secret
?
?
?
?
?Step 4 : Setting up Knowledge Base
1. Console Home>S3 > Create Bucket>upload> Add files
2. Amazon Bedrock> Knowledge Base>Create Knowledgebase> Data source>Browse S3> Add the S3 that you created in the step above > Choose an embeddings Model(Titan V2)> It uses OpenSearch serverless as vector store
3. Copy the Knowledge base ID > We need it to access Knowledge base from CodeSpaces
Step 5 : Creating Github Codespace
1.Go to Github Account>Click on the repository you want to create codespaces for > Click on code> Create codespace in main > Query your knowledge base
?
Step 6 : Cleanup & Cost for AWS resources( Make sure to delete after use):
Clean up > Make sure to delete s3 , user and codespace