AgentCore gives developers modular AWS services to build, deploy, and operate agents with any framework and model. type.com runs your team's own automations end to end, with built-in triggers, shared connections, and a thread for every run.
14-day free trial. Runs on Claude and Codex.
Updated
Amazon Bedrock AgentCore is AWS's platform for building, deploying, and operating agents with any framework and any model. It is a set of services you use together or separately: a serverless runtime that gives each session its own microVM, a managed harness, a gateway that turns APIs into MCP tools, agent identity and a token vault, memory, and CloudWatch observability. Pricing is consumption-based, starting at $0.0895 per vCPU-hour for the runtime.
type.com is a managed agent platform for a team's own work. You describe a job or write it in TypeScript, start it from a schedule, Slack, GitHub, Linear, a webhook, or email, and type.com runs it with Claude or Codex on the team's shared connections. Every run lands in a thread the team can read and continue. Plans start at $50 a month, with AI usage metered at the provider's API rates.
What decides it
Both products run agents in isolated sandboxes for you. These are the choices around that.
Nothing around the agent. Triggers, connections, delivery to Slack or email, and run history are part of the product.
The agent, from AWS's harness or your framework, plus how it starts, which tools it gets, and the interface people use.
The whole team. Runs land in shared threads that anyone in the Space can read and continue.
Developers and platform teams. AWS points business users to Amazon Quick, a separate product.
type.com's managed cloud, with no infrastructure to run, on Claude or Codex.
AWS, next to IAM, CloudWatch, and your models on Amazon Bedrock or elsewhere.
What it is
AWS describes AgentCore as an agentic platform for building, deploying, and operating agents securely at scale with any framework and foundation model. It became generally available in October 2025 and is a set of modular services: Runtime, a managed harness, Gateway, Identity, Memory, Code Interpreter, Browser, Observability, Policy, and Evaluations, among others.
There are two ways to build. With the managed harness, generally available since June 2026, you declare the model, prompt, tools, and memory in configuration, and AgentCore runs the loop on Strands Agents at no separate harness charge. Or you write the loop yourself in a framework such as LangGraph, CrewAI, Google ADK, or the OpenAI Agents SDK, and deploy it to Runtime as a zip file or container. Each session runs in its own microVM for up to 8 hours.
Triggers
An AgentCore agent starts when your code calls the InvokeAgentRuntime API, and AWS Step Functions can call it as a step in a workflow. AgentCore's documentation doesn't describe a built-in scheduler, Slack trigger, or email trigger, so a cron schedule or a Slack message reaches the agent through infrastructure you set up.
type.com starts automations from schedules (every minute to once a month), Slack messages, GitHub and Linear events, webhooks, emails sent to a Space, and RSS items. Instruction automations need no code: you describe the job, run a private test, and enable it.
Credentials
AgentCore Identity gives each agent a workload identity, and its token vault stores OAuth tokens and API keys. The gateway adds credentials to each tool call, so secrets stay out of agent code and configuration. Access is controlled with IAM, and Policy checks every tool call at the gateway against rules written in Cedar or plain language.
type.com connects apps once, through OAuth sign-in, API keys, service accounts, or hosted MCP servers, and assigns each connection to the Spaces that need it. Each run gets a short-lived token, and automation code never holds a provider credential. Scheduled runs act as the person who created the automation, with the Space's connections.
Cost
AgentCore is consumption-based with no upfront commitment. Runtime costs $0.0895 per vCPU-hour and $0.00945 per GB-hour, billed per second, the gateway costs $0.005 per 1,000 tool invocations, and memory, observability, and the other services have their own prices. Model usage is billed by the model provider, and billed usage shows up in AWS Cost Explorer.
type.com plans start at $50 a month for 2 team members, and Enterprise is $250 a month for 20. Each plan includes AI usage metered at the provider's published API rates, and team members can connect their own Claude or ChatGPT subscriptions so eligible requests run through them instead.
Full comparison
Grouped by the decisions a technical team makes when it picks where agents run.
A managed agent platform for a team's automations, with a web app, a desktop app, Slack, a CLI, and MCP.
Modular AWS services for building, deploying, and operating agents, generally available since October 2025.
Teams automating their own work, such as reports, triage, reviews, and handoffs.
Developers, and platform teams giving internal developers a paved path to production agents.
Describe the job in a Space, or write a TypeScript automation and deploy it with the type.com CLI.
Declare an agent for the managed harness, or write one in your framework and deploy it with the AgentCore CLI.
Claude or Codex, the same managed agent that works in the Space's threads.
AWS's managed harness on Strands Agents, or your own loop in LangGraph, CrewAI, Google ADK, the OpenAI Agents SDK, or another framework.
Claude and Codex models, set for each Space.
Any model, in or outside Amazon Bedrock, including Anthropic, OpenAI, Google, Amazon, Meta, and Mistral models.
An isolated Linux sandbox for shell commands, Git, the GitHub CLI, and scripts.
A dedicated microVM per session for up to 8 hours, plus Code Interpreter and Browser tools.
Connected apps, custom APIs, and hosted MCP servers, shared through the Space.
Gateway turns APIs, Lambda functions, and existing MCP servers into MCP tools.
Schedules, Slack, GitHub, Linear, webhooks, email, and RSS.
API calls from your code, including from AWS Step Functions. No built-in scheduler is documented.
A private test run in a thread only you can see. Code automations run fixtures on every deploy.
A test sandbox in the console, and Evaluations for scoring agent behavior.
A thread for every run with the agent's messages and tool calls, plus run history with outputs and errors.
Metrics, spans, and logs in Amazon CloudWatch, in an OpenTelemetry-compatible format.
Each run gets a short-lived token. Automation code never holds a provider credential.
A token vault for OAuth tokens and API keys. The gateway adds credentials to each tool call.
Public or private Spaces, private threads, and connections assigned only to the Spaces that need them.
IAM and resource-based policies, plus Policy rules in Cedar that check each tool call.
Everyone in the Space, unless a thread is private.
Whoever you grant access to the logs and traces in your AWS account.
Plans from $50 a month for 2 team members, with AI usage metered at the provider's API rates.
Consumption-based. Runtime is $0.0895 per vCPU-hour and $0.00945 per GB-hour, billed per second.
Included in the plan's AI usage, or run on a team member's own Claude or ChatGPT subscription.
Billed separately by the model provider.
AI cost by Space in the analytics dashboard, with notices at 80% and 100% of the plan's allowance.
Billed usage in AWS Cost Explorer and the Cost and Usage Report.
AgentCore is strong infrastructure, and some teams should pick it. It is probably the better fit if:
AgentCore is a set of AWS services that developers use to build, deploy, and operate agents. The alternatives are other runtimes from model and cloud providers, such as Claude Managed Agents and Microsoft Foundry Agent Service, or a managed agent platform like type.com that runs a team's automations without infrastructure to build.
| Option | Best fit |
|---|---|
| type.com | Best when a team wants its own automations run for it: triggers, shared connections, and a thread for every run, with Claude or Codex as the agent. |
| Claude Managed Agents | Fits developers building products on Claude who want Anthropic to run the agent loop and a sandbox per session, in public beta. |
| Microsoft Foundry Agent Service | Fits companies on Azure and Microsoft 365, with agents built from configuration or code, scheduled routines, and publishing to Teams and Microsoft 365 Copilot. |
| OpenAI Agents API | Fits developers on OpenAI models who want OpenAI to run a managed Codex harness, in public beta, with a sandbox OpenAI hosts or one they run. |
Summarized from each vendor's own documentation on October 15, 2026. Check them for current features and pricing.
FAQ
Yes, for automating a team's own work. AgentCore gives developers the services to build, deploy, and operate agents on AWS: a runtime, tool gateway, identity, memory, and observability. type.com is a managed agent platform where those pieces, plus triggers and a thread for every run, are already put together, with Claude or Codex as the agent.
Yes. Since June 2026, AgentCore's managed harness runs an agent you declare in configuration, with a loop built on Strands Agents and no separate harness charge. You can also write the agent loop in a framework such as LangGraph, CrewAI, Google ADK, or the OpenAI Agents SDK and deploy it to AgentCore Runtime.
Not as a built-in feature. AgentCore agents start when your code calls the InvokeAgentRuntime API, and AWS Step Functions can call them as a step in a workflow. AgentCore's documentation doesn't describe a built-in scheduler, Slack trigger, or email trigger. type.com starts automations from schedules, Slack, GitHub, Linear, webhooks, email, and RSS.
AgentCore is consumption-based with no upfront commitment: Runtime costs $0.0895 per vCPU-hour and $0.00945 per GB-hour, billed per second, with separate prices for the gateway, memory, and other services, and model usage billed by the model provider. type.com plans start at $50 a month for 2 team members and include AI usage metered at the provider's published API rates.
AgentCore works with any model, in or outside Amazon Bedrock, including models from Anthropic, OpenAI, Google, Amazon, Meta, and Mistral. type.com runs Claude and Codex, set for each Space, and team members can connect their own Claude or ChatGPT subscriptions.
Not through AgentCore itself. It is built for developers and platform teams, any interface is one you build and host, and AWS points business users to Amazon Quick, a separate product. In type.com, teammates work with the managed agent in Space threads, in Slack, and by email.
AgentCore is a set of AWS services that developers use to build, deploy, and operate agents. The alternatives are other runtimes from model and cloud providers, such as Claude Managed Agents and Microsoft Foundry Agent Service, or a managed agent platform like type.com that runs a team's automations without infrastructure to build. type.com: Best when a team wants its own automations run for it: triggers, shared connections, and a thread for every run, with Claude or Codex as the agent. Claude Managed Agents: Fits developers building products on Claude who want Anthropic to run the agent loop and a sandbox per session, in public beta. Microsoft Foundry Agent Service: Fits companies on Azure and Microsoft 365, with agents built from configuration or code, scheduled routines, and publishing to Teams and Microsoft 365 Copilot. OpenAI Agents API: Fits developers on OpenAI models who want OpenAI to run a managed Codex harness, in public beta, with a sandbox OpenAI hosts or one they run.
When your agents belong inside AWS, with IAM, CloudWatch, and the rest of your stack, when you need a model or framework type.com doesn't run, or when you are building agents into your own product and need long sessions of up to 8 hours.
type.com runs Claude and Codex automations with built-in triggers, shared connections, and a thread for every run, with no infrastructure to build around them.
Get started for free14-day free trial. Plans for 2 to 60 team members. Bring your own Claude or Codex subscription.
