Feature Request: Allow configurable model ID for mem0_memory tool #3688
Description
🚀 The feature
Currently, the mem0_memory tool in strands-tools has a hardcoded model ID (anthropic.claude-3-5-haiku-20241022-v1:0) within its DEFAULT_CONFIG. This causes a ValidationException on AWS Bedrock when this specific model is not available for on-demand use, as it requires a provisioned throughput profile.
This makes the tool inflexible and prone to breaking if the hardcoded model is deprecated or restricted.
Describe the solution you'd like I suggest modifying the mem0_memory tool to allow the LLM model ID to be configured externally. This could be achieved by:
Allowing the Mem0ServiceClient to accept an llm_model_id during initialization.
This llm_model_id would then be passed down and used to configure the Mem0 client (either OpenSearch or FAISS backend).
The mem0_memory tool function itself should also accept this llm_model_id as an optional parameter in its TOOL_SPEC.
This would allow users to specify a model that is available in their AWS region, like anthropic.claude-3-sonnet-20240229-v1:0, either directly in the tool call or via an environment variable like STRANDS_DEFAULT_MODEL_ID.
Motivation, pitch
An error occurred (ValidationException) when calling the Converse operation:
Invocation of model ID anthropic.claude-3-5-haiku-20241022-v1:0 with on-demand throughput isn’t supported.
Retry your request with the ID or ARN of an inference profile that contains this model.