Feature request: aws bedrock support

[david-wagih]

We would love to have the option to be able to use our models though aws bedrock models and authenticating with them either using AWS secrets or role id (OIDC)

[sng-asyncfunc]

https://github.com/AsyncFuncAI/deepwiki-open/pull/99/files added AWS Bedrock support branch
feature/aws-bedrock-support

Have not tried it with OIDC. Lightly tested, and require extensive review.

[FathiahHusna]

hi, may i know is deepwiki support bedrock? i tried put key_id, access_key but not working. hit this error in log.

2025-09-03 14:13:03,899 - ERROR - adalflow.components.retriever.faiss_retriever - faiss_retriever.py:291 - Error embedding queries: Model type ModelType.EMBEDDER is not supported by AWS Bedrock client
2025-09-03 14:13:03,900 - ERROR - api.rag - rag.py:487 - Error in RAG call: Model type ModelType.EMBEDDER is not supported by AWS Bedrock client
2025-09-03 14:13:03,901 - ERROR - api.websocket_wiki - websocket_wiki.py:227 - Error in RAG retrieval: 'RAGAnswer' object has no attribute 'documents'
2025-09-03 14:13:03,902 - INFO - api.rag - rag.py:73 - Dialog turns list exists but is empty
2025-09-03 14:13:03,903 - INFO - api.rag - rag.py:87 - Returning 0 dialog turns from memory
2025-09-03 14:13:03,903 - INFO - api.websocket_wiki - websocket_wiki.py:424 - No context available from RAG
2025-09-03 14:13:03,904 - ERROR - api.websocket_wiki - websocket_wiki.py:764 - Error in WebSocket handler: 'top_k'
INFO: 127.0.0.1:51457 - "GET /auth/status HTTP/1.1" 200 OK
INFO: connection closed

[Warashi]

I sent a PR to address this issue.
This PR adds AWS Bedrock support as embedding provider.
#419