choosing-an-environment changedmanaged-agents/choosing-an-environment
Nearest release: v2.1.312, published under an hour before this site recorded the change. Shown because the two are within 24 hours of each other. Nothing here says the release caused the edit.
Recorded here
Lines+14added
Lines−14removed
From line
10
where the diff opens
First seen
15 Oct 2026
this site's first read of the page
Recorded edits4to this page, all time
## What a self-hosted sandbox doesn’t change ## What a self-hosted sandbox doesn't change
The whole hunk
from line 10, old and new numbered
/
from line 10
1010 betaHeader: managed-agents-2026-04-01
1111---
1212
13An environment sets where your agent's tools and code run. They run in a sandbox. A sandbox is an isolated workspace, such as a container, where the agent runs commands and works on files. You have two options:
13An environment sets where your agent’s tools and code run. They run in a sandbox. A sandbox is an isolated workspace, such as a container, where the agent runs commands and works on files. You have two options:
1414
1515* **Cloud environment:** Anthropic runs the sandbox for you. Each session gets its own fresh Linux container on Anthropic-managed infrastructure, with common languages, databases, and utilities already installed. You choose its network rules and any extra packages to install.
16* **Self-hosted sandbox:** You run the sandbox on your own infrastructure. Claude still runs on Anthropic's side and determines which tool to call next. A program that you run, called a worker, carries out the agent's bash and file tool calls on your machines and sends each result back. It can also serve custom tools.
16* **Self-hosted sandbox:** You run the sandbox on your own infrastructure. Claude still runs on Anthropic’s side and determines which tool to call next. A program that you run, called a worker, carries out the agent’s bash and file tool calls on your machines and sends each result back. It can also serve custom tools.
1717
1818The agent definition is the same whichever you choose. You set the model, prompt, and tools on the [agent](https://platform.claude.com/docs/en/managed-agents/agent-setup), not on the environment.
1919
2020## Start with a cloud environment
2121
22If you're not sure which to choose, or you're new to Claude Managed Agents, start with a [cloud environment](https://platform.claude.com/docs/en/managed-agents/environments), as the [quickstart](https://platform.claude.com/docs/en/managed-agents/quickstart) does. Anthropic runs an isolated container for each session, with common languages and tools already installed, so there is nothing for you to host. Choose a self-hosted sandbox only if tools must run on infrastructure you control.
22If you’re not sure which to choose, or you’re new to Claude Managed Agents, start with a [cloud environment](https://platform.claude.com/docs/en/managed-agents/environments), as the [quickstart](https://platform.claude.com/docs/en/managed-agents/quickstart) does. Anthropic runs an isolated container for each session, with common languages and tools already installed, so there is nothing for you to host. Choose a self-hosted sandbox only if tools must run on infrastructure you control.
2323
2424## Compare the options
2525
from line 28
2828| | [Cloud environment](https://platform.claude.com/docs/en/managed-agents/environments) (default) | [Self-hosted sandbox](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes) |
2929| ----------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------- |
3030| Where tools and code run | An Anthropic-managed sandbox | Your worker, on your infrastructure |
31| What reaches Anthropic | The [conversation history](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#list-past-events), [tool results](https://platform.claude.com/docs/en/managed-agents/tools#available-tools), the sandbox's files and state, skills, and memory stores | The conversation history, tool results, skills, and memory stores. The sandbox's files and processes stay on your worker. |
31| What reaches Anthropic | The [conversation history](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#list-past-events), [tool results](https://platform.claude.com/docs/en/managed-agents/tools#available-tools), the sandbox’s files and state, skills, and memory stores | The conversation history, tool results, skills, and memory stores. The sandbox’s files and processes stay on your worker. |
3232| What you operate | Nothing | The worker hosts, their hardening, their network limits, and the environment key your worker authenticates with |
3333| Network reach | Network rules you set on the environment | Your own network policy |
3434| File and GitHub repository mounting | Managed by Anthropic | Managed by you |
3535| Memory stores | Mounted by Anthropic at `/mnt/memory/` | Downloaded to `/mnt/memory/` and synced by the worker |
3636| Sandbox lifecycle | Managed by Anthropic | Managed by you |
37| Choose it when | The default. You don't need tools to run on your own infrastructure. | Tools must run on infrastructure you control. |
37| Choose it when | The default. You don’t need tools to run on your own infrastructure. | Tools must run on infrastructure you control. |
3838
3939## When to choose a self-hosted sandbox
4040
4141A self-hosted sandbox fits when the agent needs to do one of the following:
4242
43* **Work on data that can't leave your network:** The agent's files and processes stay on your worker. What a tool returns to the model still reaches the Managed Agents service.
44* **Reach internal services that aren't publicly routable:** The agent's bash and file tools, and any custom tools your worker serves, run inside your network. Anthropic, not your worker, connects to the agent's MCP servers. A private one might be in front of your internal ticket system or company wiki. To reach it, you can use an [MCP tunnel](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/overview) with either option. MCP tunnels are in research preview and need an access request. See [Self-hosted sandboxes and MCP tunnels](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes#self-hosted-sandboxes-and-mcp-tunnels).
45* **Run under your organization's own compliance and audit controls:** Tools run on hosts that you configure, log, and audit.
43* **Work on data that can’t leave your network:** The agent’s files and processes stay on your worker. What a tool returns to the model still reaches the Managed Agents service.
44* **Reach internal services that aren’t publicly routable:** The agent’s bash and file tools, and any custom tools your worker serves, run inside your network. Anthropic, not your worker, connects to the agent’s MCP servers. A private one might be in front of your internal ticket system or company wiki. To reach it, you can use an [MCP tunnel](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/overview) with either option. MCP tunnels are in research preview and need an access request. See [Self-hosted sandboxes and MCP tunnels](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes#self-hosted-sandboxes-and-mcp-tunnels).
45* **Run under your organization’s own compliance and audit controls:** Tools run on hosts that you configure, log, and audit.
4646
47## What a self-hosted sandbox doesn't change
47## What a self-hosted sandbox doesn’t change
4848
49With either option, the conversation history and tool results go to the Managed Agents service, which Anthropic operates. That is how the model receives the result of each tool call. The web search and web fetch tools also run on Anthropic's side, not in the sandbox. See [Restrict web search and web fetch domains](https://platform.claude.com/docs/en/managed-agents/tools-web-restrictions). Neither option changes the compliance coverage of Managed Agents or how long it keeps conversation history. See [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility).
49With either option, the conversation history and tool results go to the Managed Agents service, which Anthropic operates. That is how the model receives the result of each tool call. The web search and web fetch tools also run on Anthropic’s side, not in the sandbox. See [Restrict web search and web fetch domains](https://platform.claude.com/docs/en/managed-agents/tools-web-restrictions). Neither option changes the compliance coverage of Managed Agents or how long it keeps conversation history. See [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility).
5050
5151## What you take on and give up with a self-hosted sandbox
5252
from line 61
6161
6262## Differences on Claude Platform on AWS
6363
64If you use Claude Managed Agents through Claude Platform on AWS, Anthropic still operates the service. There, a self-hosted sandbox can't use memory stores, workers don't use an environment key, and MCP tunnels aren't available. See [Claude Managed Agents on Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws#claude-managed-agents) and [Features not supported](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws#features-not-supported).
64If you use Claude Managed Agents through Claude Platform on AWS, Anthropic still operates the service. There, a self-hosted sandbox can’t use memory stores, workers don’t use an environment key, and MCP tunnels aren’t available. See [Claude Managed Agents on Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws#claude-managed-agents) and [Features not supported](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws#features-not-supported).
6565
6666## Differences on Claude Models on Google Cloud (Anthropic managed)
6767
68If you use Claude Managed Agents through Claude Models on Google Cloud (Anthropic managed), Anthropic still operates the service. There, memory stores aren't supported in a self-hosted sandbox, and MCP tunnels aren't available. Workers authenticate with Google Cloud credentials, not an environment key, and you write your own, because the pre-built workers in the `ant` CLI and the Python, TypeScript, and Go SDKs don't work there. See [Feature support](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed#feature-support) in the Claude Models on Google Cloud (Anthropic managed) guide.
68If you use Claude Managed Agents through Claude Models on Google Cloud (Anthropic managed), Anthropic still operates the service. There, memory stores aren’t supported in a self-hosted sandbox, and MCP tunnels aren’t available. Workers authenticate with Google Cloud credentials, not an environment key, and you write your own, because the pre-built workers in the `ant` CLI and the Python, TypeScript, and Go SDKs don’t work there. See [Feature support](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed#feature-support) in the Claude Models on Google Cloud (Anthropic managed) guide.
6969
7070## Switch from a cloud environment to a self-hosted sandbox
7171
from line 75
7575* **What is new:** A `self_hosted` environment, an environment key that you generate in the Claude Console, and a worker that you run. The [self-hosted sandboxes quickstart](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes#quickstart) covers all three.
7676* **What stops working:** The features listed in [What you take on and give up with a self-hosted sandbox](https://platform.claude.com/docs/en/managed-agents/choosing-an-environment#what-you-take-on-and-give-up-with-a-self-hosted-sandbox).
7777
78On Claude Platform on AWS, memory stores don't carry over to a self-hosted sandbox, and the worker doesn't use an environment key. See [Differences on Claude Platform on AWS](https://platform.claude.com/docs/en/managed-agents/choosing-an-environment#differences-on-claude-platform-on-aws).
78On Claude Platform on AWS, memory stores don’t carry over to a self-hosted sandbox, and the worker doesn’t use an environment key. See [Differences on Claude Platform on AWS](https://platform.claude.com/docs/en/managed-agents/choosing-an-environment#differences-on-claude-platform-on-aws).
7979
80On Claude Models on Google Cloud (Anthropic managed), memory stores aren't supported in a self-hosted sandbox, and you write your own worker, which doesn't use an environment key. See [Differences on Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/managed-agents/choosing-an-environment#differences-on-claude-models-on-google-cloud-anthropic-managed).
80On Claude Models on Google Cloud (Anthropic managed), memory stores aren’t supported in a self-hosted sandbox, and you write your own worker, which doesn’t use an environment key. See [Differences on Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/managed-agents/choosing-an-environment#differences-on-claude-models-on-google-cloud-anthropic-managed).
8181
No line in this hunk matches that.