Amazon Bedrock Agents
aws.amazon.com
Fully managed service for building generative AI agents on AWS. Configure knowledge bases, action groups (Lambda functions as tools), and guardrails through the console or API. Handles orchestration, memory, and tracing without managing infrastructure.
More in Environment, Compliance & Runtimes
AWS's open-source microVM technology powering Lambda and Fargate. Creates lightweight, KVM-based VMs in under 125ms for strong isolation. Used to build custom secure sandboxes for AI agent code execution at scale.
Deployment platform for deploying agent backends with zero configuration. Supports any language and framework with automatic deploys from Git. Offers persistent volumes and cron job scheduling for background agent processes.
Comprehensive benchmarking environment for evaluating computer-use agents on real desktop tasks. Covers web browsers, file managers, code editors, and productivity apps across Windows, macOS, and Linux — the standard eval suite for GUI-controlling agents.
AI-powered development environment where agents can write, run, and deploy code within sandboxed Repls. Natively integrates agentic code execution with cloud deployment. Ideal for demos and rapid prototyping by non-technical users.
Serverless ML platform for deploying custom AI models with cold start times under 3 seconds. Deploy any Python ML function as a scalable API endpoint. Supports continuous batching for high-throughput LLM inference.
Cloud hosting platform for deploying web services, background workers, and cron jobs with a simple configuration model. First-class support for Docker containers makes it easy to deploy agent services. Free tier suitable for prototyping.