Is my data used to train the models in Bedrock?

No. With Bedrock your data stays in your AWS environment and is not used to train the foundation models, a key point for regulated banking, healthcare or insurance companies.

Which AI models are available in Amazon Bedrock?

Bedrock offers API access to foundation models from several providers —including Anthropic (Claude), Meta (Llama), Amazon (Nova), Mistral, OpenAI and others—, with more than a hundred variants in its catalog. You can choose the model that best fits your case and switch it without rebuilding your application, all from a single managed interface.

Amazon Bedrock or building with the provider API directly?

Bedrock provides a managed layer: unified access to multiple models, your data inside your AWS environment and not used for training, RAG and agent integration, and centralized governance. Going straight to a provider API can work for a one-off prototype, but in an enterprise environment with security and compliance requirements, Bedrock's managed layer usually reduces integration and control effort.

Retrieval-Augmented Generation: a technique that connects the AI model to your knowledge base so it answers with your specific, up-to-date information rather than only what it learned during training.

Do I need to train my own model to use generative AI?

In most cases no: foundation models in Bedrock plus RAG over your data cover most enterprise needs, with far less cost and time than training a model from scratch.

← Technical glossary AWS Technologies

Amazon Bedrock — Managed GenAI

Managed access to top-tier generative AI models to build assistants, agents and automations over your own data, keeping your information under control.

What is Amazon Bedrock?

Amazon Bedrock is AWS's managed service for building generative AI applications. It gives you API access to foundation models from several providers —Anthropic, Meta, Amazon and others— without managing inference infrastructure, and lets you combine them with your corporate data securely.

With techniques like RAG (Retrieval-Augmented Generation) and agents, Bedrock connects those models to your information and systems to answer questions, automate tasks or assist your teams. The key point for regulated companies: your data is not used to train the models and stays within your AWS environment.

Use cases

What Amazon Bedrock is used for

Assistants with your knowledge

Chatbots and copilots that answer over your documentation, policies and internal data via RAG.

Agents that execute tasks

Automations that query systems and trigger actions, not just chat.

Document processing

Extract, summarize and classify large volumes of text and documents.

Support channels

Assistants over WhatsApp and other channels, integrated into your business processes.

How Caleidos implements it

Amazon Bedrock with an AWS partner

At Caleidos we take generative AI use cases to production on Bedrock —from prototype to operation— with cloud-native architecture, RAG over your data and AI agents integrated into your systems, always with compliance and control of your information. We pick the right model for each case —including GPT-5 on Amazon Bedrock and Anthropic's models— and help you decide between Claude Platform on AWS and Amazon Bedrock based on your architecture. It is our applied AI practice.

Explore Agentic AI on AWS →

Frequently asked questions

Is my data used to train the models in Bedrock?: No. With Bedrock your data stays in your AWS environment and is not used to train the foundation models, a key point for regulated banking, healthcare or insurance companies.
Which AI models are available in Amazon Bedrock?: Bedrock offers API access to foundation models from several providers —including Anthropic (Claude), Meta (Llama), Amazon (Nova), Mistral, OpenAI and others—, with more than a hundred variants in its catalog. You can choose the model that best fits your case and switch it without rebuilding your application, all from a single managed interface.
Amazon Bedrock or building with the provider API directly?: Bedrock provides a managed layer: unified access to multiple models, your data inside your AWS environment and not used for training, RAG and agent integration, and centralized governance. Going straight to a provider API can work for a one-off prototype, but in an enterprise environment with security and compliance requirements, Bedrock's managed layer usually reduces integration and control effort.
What is RAG?: Retrieval-Augmented Generation: a technique that connects the AI model to your knowledge base so it answers with your specific, up-to-date information rather than only what it learned during training.
Do I need to train my own model to use generative AI?: In most cases no: foundation models in Bedrock plus RAG over your data cover most enterprise needs, with far less cost and time than training a model from scratch.

Evaluating Amazon Bedrock for your project?

Tell us what you want to achieve. In 30 minutes we give you a concrete recommendation.

Let's talk