Our software development team uses Amazon Bedrock to access foundation models. I need to know how to access API documentation for popular large language model services that are hosted directly within the AWS ecosystem. Where are the specific API structures, IAM security policies, and Boto3 SDK parameters documented so we can securely call models like Claude, Llama, and Mistral?
3 answers
When using AWS, the documentation for these models is unified under the Amazon Bedrock User Guide on the official AWS Documentation website. You can find precise API references by searching for the Bedrock Runtime API section. This documentation outlines how to construct the InvokeModel requests using the AWS Boto3 SDK for Python or the AWS SDK for JavaScript. It also provides the JSON schema required for individual model providers, as input parameters vary significantly between a Llama model and a Claude model.
Are you structuring your codebase to use the native Bedrock Converse API endpoint, or are you writing custom payload handlers for every single individual model provider's specific JSON schema?
Make sure you review the IAM policy examples in the Bedrock documentation. Fine-grained access control is mandatory to keep your execution roles secure.
Absolutely, Kelly. Correctly configuring the resource-level permissions in your IAM policies ensures that your software development environments can only invoke the specific verified models authorized by your security team.
We are transitioning over to the Converse API endpoint because it provides a unified functional interface for multi-turn conversations. The documentation shows it handles the underlying differences beautifully, which means our software development team doesn't have to rewrite custom JSON wrappers every time we switch models.