Support for AWS Bedrock-hosted Claude 3.7 Sonnet model with the new Agent API

[uninstallit]

Validations

• I believe this is a way to improve. I'll try to join the Continue Discord for questions
• I'm not able to find an open issue that requests the same enhancement

Problem

The new Agent API currently supports Anthropic's Claude 3.7 Sonnet model when directly accessed. However, when attempting to use Claude 3.7 Sonnet via AWS Bedrock, the extension does not recognize the model and reports it as unsupported.

I expect the extension to allow me to use a supported model via AWS Bedrock.

Could you please add support for recognizing and using Anthropic Claude models hosted on AWS Bedrock within the new Agent API?

Solution

No response

[uninstallit]

Looked at the source code and I think its just the way we named the model: "bedrock-claude-v3-sonnet."

[chezsmithy]

@uninstallit Could you please share a bit more about what you are doing? Are you attempting to call AWS agents directly as a provider?

[uninstallit]

@chezsmithy I was just trying to turn on Agent mode with a Bedrock hosted Claude 3.7 Sonnet. It says not supported - but I suspect it is because the way we have named the model.

[chezsmithy]

@uninstallit yep. That would do it. Where did you overload the base model name in AWS? Just curious.

[uninstallit]

@chezsmithy Not sure - I didn't do the deployment - but I know we use another service in the middle so the model name change might not be in Bedrock.

[chezsmithy]

@uninstallit got it. Recommend changing the name if possible to match the patterns in tools support.

[kintz09]

I'm dealing with the same issue right now. Finally got all the various models I needed added to Open WebUI and my Claude 3.7 Sonnet model isn't supporting agent mode. Our Claude models also come from Bedrock but we hit our private API from Open WebUI. The model's default name is us.anthropic.claude-3-7-sonnet-20250219-v1:0 for us.

[kintz09]

For anyone else that may run into this issue, I just found the manual fix thanks to the exmaple config for Claude 4 Sonnet in the Hub.

name: Claude 4 Sonnet
version: 1.0.2
schema: v1

models:
  - name: Claude 4 Sonnet
    provider: anthropic
    model: claude-sonnet-4-20250514
    apiKey: ${{ inputs.ANTHROPIC_API_KEY }}
    roles:
      - chat
      - edit
      - apply
    defaultCompletionOptions:
      contextLength: 200000
      maxTokens: 64000
    capabilities:
      - tool_use
      - image_input

As soon as I added tool_use to my model config it immediately unlocked agent mode.

[kintz09]

For cases like these when we have to add the capabilities manually to a model's config, does anyone know how to determine whether or not a given model supports specific capabilities? I'd be nice to confirm the models' capabilities before I unleash them on my team.