Skip to content

Amazon Bedrock

Amazon Bedrock gives you access to foundation models from Anthropic, Meta, Mistral, and others — all billed through your AWS account. If you’re already using AWS, Bedrock is a natural fit.

Before connecting Selu to Bedrock, you’ll need:

Bedrock model access

Enable the models you want in the AWS console under Amazon Bedrock → Model access. Some models require a brief approval process.

IAM credentials

Create an IAM user or role with bedrock:InvokeModel permission. Generate an access key and secret key.

  1. Open the Selu dashboard and go to Settings → LLM Providers → Amazon Bedrock.
  2. Enter your AWS Access Key ID and Secret Access Key. Both are encrypted at rest.
  3. Select your AWS Region (e.g., us-east-1). Choose the region where you enabled model access.
  4. Click Test Connection to verify everything works.
  5. Save your settings.

Bedrock offers models from multiple providers. Select your model under Settings → LLM Providers → Bedrock → Model.

Popular choices include:

ModelProviderGood for
Claude SonnetAnthropicGeneral purpose, balanced speed and quality
Claude HaikuAnthropicFast, affordable responses
Llama 3MetaOpen-weight alternative, good all-rounder
Mistral LargeMistralStrong at coding and reasoning

Here’s a minimal IAM policy for the Selu user:

{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"bedrock:InvokeModel",
"bedrock:InvokeModelWithResponseStream"
],
"Resource": "arn:aws:bedrock:*::foundation-model/*"
}
]
}

Attach this policy to the IAM user whose credentials you entered in Selu.

SettingDescriptionDefault
RegionAWS region for API callsus-east-1
Max tokensMaximum response length4096
TemperatureControls randomness (0.0–1.0)0.7
  • “Access denied” — Ensure your IAM user has the correct Bedrock permissions and that the model is enabled in your selected region.
  • “Model not found” — Go to Amazon Bedrock → Model access in the AWS console and request access to the model.
  • High latency — Choose a region geographically close to your Selu server. Also consider using a faster model like Haiku.