ToolsCompaniesAbout

Amazon Bedrock AgentCore

by Amazon@amazonSeattle, Washington and Arlington, Virginia, USANo ratings yet

Managed AWS platform for building, running and governing AI agents

Visit aws.amazon.com
Tool type
Agent platform on AWS
Made for
Developers and enterprise platform teams
Runs on
AWS-managed cloud runtime with web console and CLI setup
Agents execute in AWS-managed serverless microVMs or EC2 instances in the customer's account; the console and CLI are used for setup and management.
Last checked
28 Oct 2026
3 product sources

What it does

Amazon Bedrock AgentCore is a platform for building, deploying and operating AI agents on AWS. Developers supply agent code built with frameworks such as LangGraph, Strands Agents or OpenAI Agents SDK, plus a choice of foundation model. AgentCore provides runtime hosting, tool connections through Gateway, identity and access control, memory, browser and code sandboxes, policy checks on tool calls, and observability and evaluation of agent traces.

Teams typically use it to move an agent from a prototype into production without assembling separate vendors for hosting, authentication and monitoring. Gateway can wrap existing APIs or Lambda functions as MCP-compatible tools, Policy intercepts tool calls against rules, and Observability traces each step of an agent run. Pricing is consumption-based with no upfront commitment, so each service is billed by what it uses.

It suits engineering teams building agents on AWS who need governance, identity integration and compliance alignment. Access is through the AWS console, the AgentCore CLI and APIs, and it requires an AWS account. Costs can be harder to predict because many features bill per second, per GB or per request across several meters, and some components incur separate Bedrock, CloudWatch or S3 charges.

What it can do

  • Serverless agent runtime with session isolation for deploying agents and tools
  • Gateway that converts APIs, Lambda functions and OpenAPI specs into MCP-compatible tools
  • Identity and OAuth token management for agents accessing AWS and third-party resources
  • Short-term and long-term agent memory with configurable strategies
  • Isolated browser and code interpreter sandboxes for agent actions
  • Policy checks that evaluate each tool call against defined rules before execution

Works with

  • Amazon Bedrock models
  • OpenAI, Google Gemini, Anthropic Claude, Meta Llama and Mistral models via Runtime
  • LangGraph, LangChain, Strands Agents, LlamaIndex, CrewAI and OpenAI Agents SDK frameworks
  • Playwright and BrowserUse browser automation
  • Amazon CloudWatch for observability data
  • Coinbase CDP and Stripe (Privy) wallet providers for the Payments service

Example workflows

Hosting a tool-using support agent
Input: agent code built with Strands or LangGraph and an existing internal REST API. Steps: wrap the API as an MCP tool through Gateway, deploy the agent to Runtime, attach Policy rules limiting which actions it may call, and review traces in Observability. Output: a hosted agent whose tool calls are authorized and logged.
Tuning agent prompts from production traces
Input: traces from a running agent instrumented with OpenTelemetry. Steps: run Evaluations on sessions, then use optimization to generate system prompt or tool description changes and test them with traffic splitting through Gateway. Output: a versioned configuration candidate compared against the current one in an A/B experiment.

Before you choose

  • Cost is metered across many components (microVM CPU and memory, Gateway operations, Policy authorization requests, Memory events, CloudWatch telemetry), so estimates require the pricing calculator or per-service rate tables.
  • Several services depend on other AWS services, including Bedrock model charges, S3 storage for browser profiles, and CloudWatch for observability data.
  • Some features are in preview or are scheduled to launch later, such as managed session storage (public preview) and committed-baseline runtime pricing (launching by October 2026).

Listed under

Get started

Create an AWS account, then follow the CLI getting-started guide.

  1. Create or sign in to an AWS account and open the AgentCore documentation.
  2. Install and configure the AgentCore CLI following the get-started guide.
  3. Deploy your agent to Runtime, then connect tools through Gateway and set up Identity for authentication.
Sign up and install the AgentCore CLI

Plans and prices

Prices in USD

New AWS customers receive up to $200 in Free Tier credits; the page does not specify terms or whether these apply to AgentCore usage beyond that statement. Runtime microVMs bill per second on CPU and memory with a 128MB minimum; Instances bill EC2 cost plus a management fee. Web Search is $7 per 1,000 queries. Gateway data egress to customer VPCs is $0.006 per GB. Use the pricing calculator and per-service rate tables for estimates.

Usage-based (per component)$7 per 1,000 Web Search queriesRuntime microVMs; 1-second minimum, 128MB minimum memory. Check the pricing page rate table for current rates.Web Search, as stated on the pricing page.
  • No upfront commitments or minimum fees
  • Each feature billed independently or together

Source: aws.amazon.com/bedrock/agentcore/pricing, checked 28 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.

Ratings and reviews

No ratings for Amazon Bedrock AgentCore yet.

Be the first to rate it

What’s changed

Free account

See what’s changed at Amazon Bedrock AgentCore

Launches, price moves and pivots, each with a date and a source. Free with an account.

Signals

Free account

See Amazon Bedrock AgentCore’s signals

Launch traction, hiring and release activity, tracked over time. Free with an account.

Data and security

Free account

See how Amazon Bedrock AgentCore handles your data

Training on your files, compliance, hosting and the models behind it. Free with an account.

Compare with similar tools

Free account

Compare Amazon Bedrock AgentCore side by side

Prices, platforms and features against similar tools, with sources. Free with an account.

Where this comes from

How we research
Company website
aws.amazon.com: product, pricing and policy pages
YC directory mirror
Batch and company name
Web search
Confirms the official site
Social accounts
Accounts the company links, and its posts on X
Human review
Checked by a person
knowledge_base

Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.