Build
To start building models using Amazon Bedrock, first, start from your use-case. Then choose an API, an endpoint, and then start using the models programmatically.
Choose API based on your use-case
Amazon Bedrock provide four main API patterns to perform inference in Amazon Bedrock. Responses, Chat Completions, Invoke, and Converse. Read more about the APIs supported.
| Scenario | Recommended API |
|---|---|
| Migrating from OpenAI API-compatible endpoint | Use OpenAI-compatible APIs: Responses API |
| Using models not compatible with OpenAI-compatible endpoint | Use native Amazon Bedrock APIs: Converse and Invoke. Read more on choosing the right APIs for your use-case. |
Choosing end-point for Amazon Bedrock
Once you have identified which API to use, you can then identify the endpoint to use to interact programmatically with Amazon Bedrock. Read more about the APIs supported.
| Endpoint | Supported APIs |
|---|---|
bedrock-mantle.* |
Responses API, Chat Completions API |
bedrock-runtime.* |
Invoke API, Converse API, Chat Completions API |
Choosing a model
You can see the list of models that Amazon Bedrock supports based on the endpoint you are using.
| Endpoint | API | Description |
|---|---|---|
bedrock-runtime.* |
ListFoundationModels |
Returns a list of all available foundation models in Bedrock. Provides summary information about multiple models, including their model IDs, provider names, supported modalities (text, image, embedding), input/output formats, and whether they support streaming or customization. Use GetFoundationModel API to retrieve detailed information about a specific foundation model using its model ID. |
bedrock-mantle.* |
client.models.list |
OpenAI-compatible API to discover available models. Retrieves a list of models you can use with the Responses API and Chat Completions API. |
