---
title: Amazon Bedrock AgentCore Consulting
description: Production agent infrastructure on Amazon Bedrock AgentCore — Runtime, Gateway, Memory, Identity, Observability and Evaluations. Bring your own framework; we operate the substrate.
url: https://www.factualminds.com/services/amazon-bedrock-agentcore/
category: agents
updated: 2026-08-30
---

# Amazon Bedrock AgentCore Consulting

> The managed substrate your agents run on — isolated runtime, Gateway for tool access, Memory with fine-grained access control, identity propagation, observability and Evaluations. Bring Strands, LangGraph, or a plain control loop.

## What AgentCore actually gives you

Every team that ships an agent ends up building the same six things: somewhere isolated to run it, a governed way to reach tools, memory that survives the session, an identity that travels with the request, traces you can read when it misbehaves, and a way to tell whether it is any good.

You can build all six. Teams do. The problem is that you then own them — forever, alongside the product you actually meant to build.

Amazon Bedrock AgentCore provides them as managed services: Runtime, Gateway, Memory, Identity, Observability and Evaluations, plus a Managed Harness and optional Payments. It is bring-your-own-framework, so Strands, LangGraph, or three sequential prompts with an explicit control loop all sit on top of it unchanged.

## The lifecycle change worth knowing about

On June 30, 2026 AWS renamed the original Bedrock Agents — the November 2023 service — to **Bedrock Agents Classic**. It entered maintenance for new customers after **July 30, 2026**.

Existing Classic deployments continue to operate, so this is not an emergency. But it does mean two things. Net-new agent builds should start on AgentCore. And teams currently running Classic agents should have a deliberate plan rather than discovering the situation during an incident.

## Where the boundary lives

The most common production failure in agent systems is not a bad model output. It is an agent with more write access than anyone intended, discovered after it used it.

Gateway is the answer to that. Every tool an agent can reach passes through it, with customer-configurable rate limits and Cedar-based fine-grained access control on writes — extended to Memory namespaces on August 28, 2026. The practical consequence: mutating an order, a price, or a customer record becomes a permission decision evaluated outside the model. A model can be argued out of a system prompt. It cannot be argued out of an authorization policy.

Identity propagation is the other half. When the acting user identity travels with the request, instead of every tool call arriving as one over-privileged service role, the audit trail becomes something a compliance reviewer can actually use.

## Evals are not optional

An agent without an eval suite is a system whose quality you are asserting rather than measuring.

We build a golden dataset per agent during development and set a pass bar it must clear before deployment. AgentCore Evaluations runs it on every deploy. If it does not clear, it does not ship — the sprint calendar does not get a vote. This is the single practice that most reliably separates agents that are still running in ninety days from agents that were quietly switched off.

## Modelling the bill honestly

AgentCore bills across several components rather than one meter, and first estimates are usually wrong in the same direction — too low, because the model inference line dominates and tool-call volume per conversation is underestimated.

We publish a free [Amazon Bedrock AgentCore pricing calculator](/tools/amazon-bedrock-agentcore-pricing-calculator/) covering the components. It is the same one we use on the first call, and we would rather you ran it before talking to us.

## What we do

We design the AgentCore stack against your actual workflow, deploy it as infrastructure-as-code in your accounts, wire the Gateway boundary and Cedar policies, build the eval harness and the golden dataset, instrument cost ceilings and CloudWatch alarms, and hand over the runbooks.

Your engineers are in the pull requests throughout. There is no proprietary orchestration layer to be locked into — the whole point of AgentCore is that the substrate is AWS-managed and the agent logic is yours.

For the commerce-specific application of all of this, see [eCommerce AI Agents on AWS](/services/ecommerce-ai-agents/). For the broader Bedrock platform — Knowledge Bases, model selection, Guardrails — see [Amazon Bedrock Consulting](/services/aws-bedrock/).

## Features

### AgentCore Runtime
Isolated execution per agent session, so one runaway loop cannot take a neighbour down. Serverless scaling with the session lifecycle handled for you rather than reimplemented per project.

### Gateway and Tool Access
One governed entry point for every tool an agent can reach, with customer-configurable rate limits and Cedar-based fine-grained access control on writes. The tool catalog stops being a convention and becomes an enforced boundary.

### Memory With Access Control
Session and long-term memory with flexible namespaces and fine-grained access control, so a multi-tenant agent cannot read across tenant boundaries by accident.

### Identity Propagation
The acting user identity travels with the request instead of every tool call arriving as one over-privileged service role. This is what makes an agent audit trail worth reading.

### Observability and Evaluations
Traces per agent turn, and a golden dataset with a pass bar the agent must clear on every deployment. If it does not clear, it does not ship.

### Cost Instrumentation
Per-conversation token budgets, model routing by task, and CloudWatch alarms that fire before the threshold. AgentCore has a multi-component bill; we model it before you commit.

## Why FactualMinds

### Bring Your Own Framework
AgentCore is framework-agnostic. Strands, LangGraph, or three sequential prompts with an explicit control loop all run on it. We have no incentive to sell you an abstraction you will later have to debug.

### The Migration Path Is Real
Bedrock Agents Classic entered maintenance for new customers after July 30, 2026. Existing deployments keep running, but net-new builds should start on AgentCore — and teams with Classic agents need a plan rather than a surprise.

### We Model the Bill Before the Build
AgentCore prices across multiple components, and the naive estimate is usually wrong in the same direction. Our free pricing calculator is the one we use on the first call.

### Minimal Abstraction, Debuggable Code
Framework-heavy agent stacks are hard to debug at 2am. We keep the control flow explicit and readable, because you are the one who will be operating it.

## FAQ

### What is the difference between Bedrock Agents Classic and AgentCore?
On June 30, 2026 AWS renamed the original Bedrock Agents to Bedrock Agents Classic, and it entered maintenance for new customers after July 30, 2026. AgentCore is the current platform: Runtime, Gateway, Memory, Identity, Observability, Managed Harness, Evaluations and optional Payments. Existing Agents Classic deployments continue to operate — but net-new agent builds should start on AgentCore.

### Do we still need LangGraph or Strands?
Only if the orchestration genuinely needs them. AgentCore is bring-your-own-framework, so both run on it, as does a plain control loop with explicit sequential prompts. Our default is the least abstraction that solves the problem, because whatever we build is what your team debugs later. Framework choice is a decision we make with you against the actual workflow, not a house preference.

### What is AgentCore Payments?
It reached GA on August 18, 2026 and covers agent-to-API microtransactions against a session budget — an agent paying for an API call, an MCP server, or another agent. It is not a consumer checkout replacement. If you are looking at shopping agents buying products from your store, that is the Agentic Commerce Protocol and Universal Commerce Protocol territory, covered by our Agentic Commerce Readiness engagement.

### How do we control what an agent is allowed to do?
Gateway is where the boundary lives. Every tool an agent can reach passes through it, with customer-configurable rate limits and Cedar-based fine-grained access control on writes — added to Memory namespaces on August 28, 2026. That makes a write a permission decision evaluated outside the model, rather than an instruction the model could be argued out of. Bedrock Guardrails handle content-level filtering on top.

### What does AgentCore cost?
It bills across several components rather than a single meter, which is why first estimates are usually wrong. The dominant costs are typically model inference and runtime session time; the rest depends on memory retention, gateway call volume and observability retention. We publish a free AgentCore pricing calculator covering the components so you can model your own workload before committing to a build.

---

*Source: https://www.factualminds.com/services/amazon-bedrock-agentcore/*