-
Couldn't load subscription status.
- Fork 6.5k
Description
Feature Description
In a recent update, support for the global cross-region inference profile prefix was added to llama-index-llms-bedrock-converse (see PR 20064).
Similarly, AWS now provides a "jp" (Japan) cross-region inference profile for Bedrock Claude Sonnet 4.5 and Bedrock Claude Haiku 4.5 models.
I would like to request that llama-index-llms-bedrock-converse support the "jp" cross-region inference profile for these models, allowing users to select the Japan profile as needed.
Reference:
- AWS official documentation for supported cross-region inference profiles: https://docs.aws.amazon.com/bedrock/latest/userguide/inference-profiles-support.html
Reason
Currently, while the global cross-region inference profile is supported, there is no way to select or utilize the "jp" (Japan) cross-region profile, which can route requests specifically via the Japanese region's optimized resources.
This limitation is problematic for users and workloads that benefit from region-specific inference, such as compliance requirements, latency optimization, or data residency concerns specific to Japan.
Value of Feature
Adding support for the "jp" cross-region inference profile would:
- Enable users to leverage lower latency and regional optimization for workloads specific to Japan.
- Increase compliance with data residency requirements for Japanese customers.