Concepts

AWS Bedrock

AWS managed service providing access to foundation models from multiple providers (Anthropic, Meta, Mistral) via API, without managing ML infrastructure.

seed#aws#bedrock#llm#ai#foundation-models#serverless

What it is

Amazon Bedrock is a serverless service providing access to foundation models from multiple providers through a unified API. No infrastructure to manage — just call the API and pay per tokens consumed.

Available models

ProviderModels
AnthropicClaude 3.5 Sonnet, Claude 3 Opus/Haiku
MetaLlama 3.1, Llama 3.2
MistralMistral Large, Mixtral
AmazonTitan Text, Titan Embeddings
CohereCommand R, Embed
Stability AIStable Diffusion

Features

FeatureFunctionIntegration
Knowledge BasesManaged RAGS3 as source, OpenSearch as vector store
AgentsAI agents with tools and memoryLambda functions as tools
GuardrailsContent and PII filtersApplicable to any model
Fine-tuningModel customizationOwn data in S3
Model evaluationCompare models on your dataAutomatic + human metrics

Agent integration

Bedrock Agents allows creating AI agents that use tools (Lambda functions) and query knowledge bases, all managed.

Why it matters

Bedrock democratizes access to foundation models without the complexity of managing GPU infrastructure. For teams already operating on AWS, it is the most direct way to integrate generative AI into existing applications, with the security and compliance of the AWS ecosystem.

References

Concepts