Bedrock Model Access

Required models

The demo uses two models:

Purpose	Model ID	Output
Embeddings	`amazon.titan-embed-text-v2:0`	1024-dimensional float32 vector
LLM (answer generation)	`au.anthropic.claude-sonnet-4-5-20250929-v1:0`	Structured markdown text

Enable model access

Open the Amazon Bedrock console
Select your deployment region
Navigate to Model access in the left sidebar
Click Manage model access
Enable:
- Amazon → Titan Text Embeddings V2
- Anthropic → Claude Sonnet 4.5 (or your chosen model)
Click Request model access
Wait for status to show Access granted

Verify access

./scripts/check-bedrock-access.sh

Or manually:

aws bedrock get-foundation-model \
  --model-identifier amazon.titan-embed-text-v2:0 \
  --region ap-southeast-2 \
  --query 'modelDetails.modelId' \
  --output text

Cross-region inference profiles

The LLM model ID (au.anthropic.claude-sonnet-4-5-20250929-v1:0) is an AU cross-region inference profile. This routes requests to the nearest available region (ap-southeast-2 or ap-southeast-4) for lower latency.

IAM permissions must cover both:

The inference profile ARN: arn:aws:bedrock:{region}:{account}:inference-profile/{profile_id}
The underlying foundation model ARNs for each routing region