Enhance Bedrock Provider Configuration and Header Management

- Added `forward_client_headers_to_llm_api` setting in the Bedrock documentation to facilitate client-side header forwarding.
- Updated `completion` function to use merged headers instead of original `extra_headers`.
- Improved request handling in `BedrockConverseLLM` and `AmazonInvokeConfig` to ensure proper header management for `anthropic-beta` parameters.
- Refactored request transformation logic to return the transformed request for better clarity and functionality.

docs/my-website/docs/providers/bedrock.md

@@ -687,6 +687,9 @@ model_list:
       model: bedrock/converse/anthropic.claude-3-5-sonnet-20241022-v2:0
       extra_headers:
         anthropic-beta: "computer-use-2024-10-22,context-1m-2025-08-07"
+
+general_settings:
+  forward_client_headers_to_llm_api: true  # 👈 Required for client-side header forwarding
 ```
 
 **Set on Request**
@@ -711,6 +714,10 @@ response = client.chat.completions.create(
 )
 ```
 
+:::info
+**For client-side header forwarding**: When using the proxy and sending `anthropic-beta` headers from the client (like the OpenAI SDK), you need to enable `forward_client_headers_to_llm_api: true` in your proxy's `general_settings`. This tells the proxy to extract headers from HTTP requests and forward them to the underlying LLM provider.
+:::
+
 </TabItem>
 </Tabs>
 

litellm/llms/bedrock/chat/converse_handler.py

@@ -189,6 +189,7 @@ class BedrockConverseLLM(BaseAWSLLM):
             headers=headers,
         )
         data = json.dumps(request_data)
+        
         prepped = self.get_request_headers(
             credentials=credentials,
             aws_region_name=litellm_params.get("aws_region_name") or "us-west-2",
@@ -395,6 +396,7 @@ class BedrockConverseLLM(BaseAWSLLM):
             headers=extra_headers,
         )
         data = json.dumps(_data)
+        
         prepped = self.get_request_headers(
             credentials=credentials,
             aws_region_name=aws_region_name,

litellm/llms/bedrock/chat/invoke_transformations/base_invoke_transformation.py

@@ -190,13 +190,15 @@ class AmazonInvokeConfig(BaseConfig, BaseAWSLLM):
                     ] = True  # cohere requires stream = True in inference params
                 request_data = {"prompt": prompt, **inference_params}
         elif provider == "anthropic":
-            return litellm.AmazonAnthropicClaudeConfig().transform_request(
+            transformed_request = litellm.AmazonAnthropicClaudeConfig().transform_request(
                 model=model,
                 messages=messages,
                 optional_params=optional_params,
                 litellm_params=litellm_params,
                 headers=headers,
             )
+            
+            return transformed_request
         elif provider == "nova":
             return litellm.AmazonInvokeNovaConfig().transform_request(
                 model=model,

litellm/main.py

@@ -2981,7 +2981,7 @@ def completion(  # type: ignore # noqa: PLR0915
                     logger_fn=logger_fn,
                     encoding=encoding,
                     logging_obj=logging,
-                    extra_headers=extra_headers,
+                    extra_headers=headers,  # Use merged headers instead of original extra_headers
                     timeout=timeout,
                     acompletion=acompletion,
                     client=client,