# Claude Managed Agents

Pre-built, configurable agent harness that runs in managed infrastructure. Best for long-running tasks and asynchronous work.

Source: https://wheredoirunthis.com/products/claude-managed-agents · JSON: https://wheredoirunthis.com/api/products/claude-managed-agents · Read on 2026-10-12

## At a glance

- Type: Workflow engine
- Kind: Hosted
- Isolation: Container
- Free tier: Not documented
- Starting price: Not disclosed
- GPU: Not documented
- Regions: Not disclosed
- Coverage: 14 of 61 assessed capabilities supported
- Agent readiness: read 10 of 11 signals, drive 10 of 11 signals

## Links

- Homepage: https://claude.com/platform/api
- Docs: https://platform.claude.com/docs/en/managed-agents/overview
- Pricing: https://platform.claude.com/docs/en/about-claude/pricing.md

## Pricing

- Model: Usage based: model tokens plus running session time; server-side web search separately billed.
- Unit: Session-hour at $0.08, measured to millisecond while running; token rates vary by model/cache/speed/inference geography; web search $10/1,000 searches.
- Free tier: New users receive a small amount of free credits to test the API; amount not quantified. No ongoing free Managed Agents tier advertised.

| Tier | Price | Includes | Notes |
| --- | --- | --- | --- |
| Managed Agents beta | $0.08 per running session-hour + model tokens + applicable tools | Enabled by default for all API accounts; MCP tunnels and dreaming separately require research-preview access. | No disclosed fixed subscription/included session-hour allowance. Research-preview access is not a general sales gate. [source](https://platform.claude.com/docs/en/about-claude/pricing.md) |

- Overages: Model token table applies. US-only inference 1.1x on supported models; fast-mode premium and prompt cache multipliers apply. No separate code-execution container-hour charge on top of session runtime.
- Card required: Not documented
- Source: https://platform.claude.com/docs/en/about-claude/pricing.md

## Runtime

- **GPU & accelerators**: Not documented [source](https://platform.claude.com/docs/en/managed-agents/cloud-sandboxes-reference.md) (confidence: low)
- **BYOC (Bring your own cloud)**: Optional self-hosted sandbox workers execute tools in customer infrastructure; agent harness/control plane, session/work queue and memory remain Anthropic-hosted. [source](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes.md) (confidence: high)
- **Secrets management**: Workspace-scoped per-user vault references; platform proxy handles external MCP tokens. Any API key with workspace access can reference vaults. Worker environment key scopes to work queue; per-session secret required for memory. [source](https://platform.claude.com/docs/en/managed-agents/vaults.md) (confidence: high)
- **Storage options**: Cloud sandbox disk up to 10 GB; memory up to 8 GB. Files API uploads can be mounted; session files downloadable. Memory stores hold cross-session text documents with version history. [source](https://platform.claude.com/docs/en/managed-agents/cloud-sandboxes-reference.md) (confidence: high)
- **Supported workloads**: Long-running asynchronous coding/research/data/document tasks, bash/file/Playwright tools in Linux containers, Remote MCP and custom tool integrations, Recurring scheduled agents and signed webhook triggers, Subagents, advisor model and dynamic workflow runs, Outcome-graded iteration, Cross-session memory; gated dreaming consolidation [source](https://platform.claude.com/docs/en/managed-agents/overview.md) (confidence: high)
- **Persistence**: Stateful sessions retain conversation log, sandbox state and outputs across pauses; session deletion permanently deletes scoped files. Cross-session memory stores remain hosted, with text-document version history. [source](https://platform.claude.com/docs/en/managed-agents/overview.md) (confidence: high)
- **Data jurisdiction**: US/global model inference routing can be pinned by agent model config. Physical sandbox and session-storage region/residency are not disclosed here. [source](https://platform.claude.com/docs/en/manage-claude/data-residency.md) (confidence: medium)
- **Resource limits & quotas**: Create endpoints 300/minute; read/list/stream endpoints 1,200/minute per organization; shared organization spend/rate limits apply. Cloud memory up to 8 GB, disk up to 10 GB. Memory stores cap at 10,000 memories. vCPU sizing and global maximum concurrent sessions not established by these pages. [source](https://platform.claude.com/docs/en/managed-agents/reference.md) (confidence: high)
- **SDK languages**: Python, TypeScript, C#, Go, Java, PHP, Ruby [source](https://platform.claude.com/docs/en/cli-sdks-libraries/overview.md) (confidence: high)
- **Network & egress controls**: unrestricted, limited domain allowlists/package-manager/MCP allowances, or none; limited rules also constrain server-side web_search/web_fetch. API-created cloud default unrestricted; Studio-created default limited. [source](https://platform.claude.com/docs/en/managed-agents/environments.md) (confidence: high)
- **Snapshots & forks**: Not documented [source](https://platform.claude.com/docs/en/managed-agents/memory.md) (confidence: low)
- **Deployment methods**: REST API, official ant CLI, official client SDK beta methods, Console visual builder [source](https://platform.claude.com/docs/en/managed-agents/onboarding.md) (confidence: high)
- **Isolation**: Container [source](https://platform.claude.com/docs/en/managed-agents/environments.md) (confidence: high)

## Capabilities

### AI & agents

| Capability | Status | Evidence |
| --- | --- | --- |
| Agent skills | Supported | "Attach Anthropic or custom skills from your organization's library." [source](https://platform.claude.com/docs/en/managed-agents/onboarding.md) |
| AI app builder | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Autonomous QA agent | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Built-in coding agent | Supported | "Pre-built, configurable agent harness that runs in managed infrastructure. Best for long-running tasks and asynchronous work." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| BYOK for LLMs | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Credential-injection proxy | Supported | "Claude calls MCP tools via a dedicated proxy; this proxy takes in a token associated with the session." [source](https://www.anthropic.com/engineering/managed-agents) |
| LLM gateway | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| MCP client | Supported | "MCP servers" [source](https://platform.claude.com/docs/en/managed-agents/agent-setup.md) |
| MCP-server hosting | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| MCP server | Partially supported | "Search and read the Claude documentation and API reference." [source](https://platform.claude.com/docs/mcp/server-card) |
| Model inference API (first-party) | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Preinstalled coding agent | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |

### Networking & delivery

| Capability | Status | Evidence |
| --- | --- | --- |
| Automatic HTTPS | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Custom domains | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Direct public IPv6 | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Edge KV store | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Egress controls | Supported | "each session gets its own isolated sandbox (a fresh Linux container)." [source](https://platform.claude.com/docs/en/managed-agents/environments.md) |
| Email send / receive | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Public preview URLs | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Private networking | Not supported | "wrap an MCP server inside your network as custom tools without running a tunnel." [source](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes-custom-tools.md) |
| Built-in reverse proxy | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |

### Organisation & security

| Capability | Status | Evidence |
| --- | --- | --- |
| Built-in auth | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| BYOC | Partially supported | "title: Self-hosted sandboxes" [source](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes.md) |
| Deploy into customers' VPCs | Not supported | "Because it runs inside your sandbox, the tool reaches the internal services, credentials, and network egress you configured for the sandbox, and nothing more." [source](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes-custom-tools.md) |
| Secrets management | Supported | "title: Authenticate with vaults" [source](https://platform.claude.com/docs/en/managed-agents/vaults.md) |
| Self-host / OSS core | Not supported | "Anthropic secures the control plane across all environments: session and work queue integrity, multitenant isolation, and agent-context minimization." [source](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes-security.md) |
| Sharing | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| SSO / SAML | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Teams | Supported | "The session viewer is only accessible to Developers and Admins." [source](https://platform.claude.com/docs/en/managed-agents/session-observability.md) |
| Workload identity | Not supported | "The workload exchanges its IdP-issued JWT at `POST /v1/oauth/token` for a short-lived Claude API access token, and the SDK refreshes that token automatically before it expires." [source](https://platform.claude.com/docs/en/manage-claude/authentication.md) |

### Developer access

| Capability | Status | Evidence |
| --- | --- | --- |
| CLI | Supported | "ant beta:sessions create" [source](https://platform.claude.com/docs/en/managed-agents/onboarding.md) |
| E2B-compatible API | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Filesystem API | Supported | "## Listing and downloading session files" [source](https://platform.claude.com/docs/en/managed-agents/files.md) |
| Git deploy | Not supported | "For Git, we use each repository's access token to clone the repo during sandbox initialization and wire it into the local git remote." [source](https://www.anthropic.com/engineering/managed-agents) |
| HTTP data API | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| OpenAPI spec | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Private container registry | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Public API | Supported | "https://api.anthropic.com/v1/sessions" [source](https://platform.claude.com/docs/en/managed-agents/onboarding.md) |
| SDK | Supported | "```python Python" [source](https://platform.claude.com/docs/en/managed-agents/onboarding.md) |
| SSH access | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |

### Compute & storage

| Capability | Status | Evidence |
| --- | --- | --- |
| Cloud-init / ignition config | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Desktop / computer-use sandbox | Not supported | "Playwright (Python and Node.js) - Browser automation library" [source](https://platform.claude.com/docs/en/managed-agents/cloud-sandboxes-reference.md) |
| GPU compute | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Live-state (memory) snapshots | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Multi-region | Not supported | "The `inference_geo` parameter controls where model inference runs for a specific API request." [source](https://platform.claude.com/docs/en/manage-claude/data-residency.md) |
| Object storage | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Persistent disk | Partially supported | "sessions are long-running, resume cleanly after pauses, and store conversation history, sandbox state, and outputs server-side." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Sandbox templates / custom images | Not supported | "You create an environment once, then reference its ID each time you start a session." [source](https://platform.claude.com/docs/en/managed-agents/environments.md) |
| Scale to zero | Supported | "Runtime is measured to the millisecond and accrues only while the session's status is `running`." [source](https://platform.claude.com/docs/en/about-claude/pricing.md) |
| Selectable isolation runtime | Not supported | "Not advertised on the public Managed Agents overview, runtime guides, integration docs, pricing, and API surfaces captured in this scan. This is absence of public evidence, not proof of impossibility." [source](https://platform.claude.com/docs/en/managed-agents/overview) |
| Snapshots / forking | Not supported | "Memory stores let the agent carry information across sessions: user preferences, project conventions, prior mistakes, and domain context." [source](https://platform.claude.com/docs/en/managed-agents/memory.md) |

### Data

| Capability | Status | Evidence |
| --- | --- | --- |
| Database branching | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Ephemeral instant DB (no-signup) | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| First-party database | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Vector search | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |

### Platform & operations

| Capability | Status | Evidence |
| --- | --- | --- |
| Managed CI/CD runners | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Built-in observability | Supported | "The list shows every session in the workspace with its status, agent, token usage, cost, and creation time." [source](https://platform.claude.com/docs/en/managed-agents/session-observability.md) |
| Preview environments | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Scheduled tasks | Supported | "title: Scheduled deployments" [source](https://platform.claude.com/docs/en/managed-agents/scheduled-deployments.md) |
| Project template marketplace | Not applicable | "Pre-built, configurable agent harness that runs in managed infrastructure." [source](https://platform.claude.com/docs/en/managed-agents/overview.md) |
| Durable workflows | Partially supported | "title: Session operations" [source](https://platform.claude.com/docs/en/managed-agents/session-operations) |

## Agent readiness

### Read: 10 of 11 signals present

10/11 signals (7 full, 3 partial)

- Llms txt: pass — "https://platform.claude.com/llms.txt: substantive 82,905-byte documentation index, including Managed Agents and pricing."
- Llms full txt: pass — "https://platform.claude.com/llms-full.txt: 43,216,345 bytes HTTP 200; comprehensive index/content inspected, retained pointer per artifact cap."
- Robots AI policy: partial — "https://platform.claude.com/robots.txt: User-Agent: *; Disallow: /api; no explicit AI-UA policy."
- JSON LD: fail — "Subject HTML has no application/ld+json script; web-surface-signals.json records the observed head."
- Sitemap: pass — "https://platform.claude.com/sitemap.xml: 372,340-byte XML urlset; pointer retains multiple entries."
- Meta quality: pass — "Descriptive title/description, OG URL/image/title/description and Twitter cards on the Managed Agents overview."
- Markdown negotiation: pass — "GET subject with Accept: text/markdown returned HTTP 200, Content-Type: text/markdown; charset=utf-8; .md alternate works."
- Content efficiency: partial — "Subject HTML: 7550/334843 readable bytes = 2.25%; 1–5% ladder band."
- Docs discoverable: pass — "https://claude.com/platform/api links directly to the Managed Agents overview and memory docs; header has Developer docs."
- Pricing machine readable: pass — "Pricing markdown has structured rate tables: session runtime $0.08/session-hour, tokens by model and worked examples."
- Anti bot: partial — "Public overview/docs/marketing fetched normally; favicon.svg returned 403 to plain fetch, recovered as actual SVG via sequential foxcrawl. Trust portal required JS rendering."

### Drive: 10 of 11 signals present

10/11 signals (9 full, 1 partial)

- Public API: pass — "Documented REST agents/sessions/environments/events/memory/workflows endpoints and beta client methods."
- Openapi spec: fail — "API intro/auth provides navigable references, not a downloadable spec link. /openapi.json, /docs/openapi.json and /swagger.json on platform plus /openapi.json and /v1/openapi.json on api.anthropic.com returned 404."
- MCP server: pass — "https://platform.claude.com/docs/mcp documented via valid server-card JSON: search/read docs MCP. Not an agent control-plane server."
- SDK packages: pass — "Official client SDKs in Python, TypeScript, C#, Go, Java, PHP and Ruby with Managed Agents beta methods."
- Auth docs: pass — "API-key/Bearer/workspace scope docs, OAuth token exchange via Workload Identity Federation; environment-scoped worker keys and per-session secrets documented separately."
- Agent self serve creds: pass — "Workload Identity Federation documents POST /v1/oauth/token for programmatic short-lived Claude API tokens; service account/issuer/rule setup prerequisite. API-key creation itself remains a Console step; no signup attempt."
- Webhooks: pass — "Managed Agents webhook guide documents endpoint registration, session event types and signature verification."
- CLI tool: pass — "Official ant CLI, anthropics/anthropic-cli; beta:sessions create and worker start documented."
- Error model: pass — "API errors document JSON type:error, nested error type/message and request_id, with HTTP error classes."
- Rate limit docs: pass — "Managed Agents reference: 300 create/minute, 1,200 read/minute per organization; shared API rate-limit headers documented."
- Agent config: partial — "Public managed-agent skills guide/libraries expose a skills listing surface; /AGENTS.md and /.well-known/agent returned HTTP 404. No multiple agent-instruction discovery surfaces established."

## Compliance & regions

- SOC 2 Type II: certified (scope: hosted) — "Claude via Anthropic's API | SOC 2 Type 2 ✅" [source](https://trust.anthropic.com/)
- ISO/IEC 27001: certified (scope: hosted) — "Claude via Anthropic's API | ISO 27001 ✅" [source](https://trust.anthropic.com/)
- CSA STAR: self-attested (scope: hosted) — "Claude via Anthropic's API | CSA Star ✅" [source](https://trust.anthropic.com/)
- Note: Rendered trust center public matrix observed. API-wide attestations above are medium-confidence scope, not gated service-specific reports. Managed Agents overview explicitly says NOT eligible for Zero Data Retention or HIPAA BAA: the broader API HIPAA checkmark MUST NOT be inherited. ISO 42001 and NIST 800-171 appear in vendor table but are outside current compliance registry; kept here without admitting keys. FedRAMP/DoD apply only to other deployment SKUs, not this service.
- Regions note: Model inference geography supports global/US routing; sandbox/session storage regions or a regional availability map are not disclosed. Do not infer service regions from model routing or AWS sibling distribution.

## Company & presence

- Privacy policy: https://www.anthropic.com/legal/privacy
- Terms of service: https://www.anthropic.com/legal/commercial-terms
- Marketing: https://claude.com/platform/api
- Managed Agents docs: https://platform.claude.com/docs/en/managed-agents/overview
- Pricing: https://platform.claude.com/docs/en/about-claude/pricing.md
- Console: https://platform.claude.com/
- API client CLI repo: https://github.com/anthropics/anthropic-cli
- Managed Agents example repo: https://github.com/anthropics/claude-quickstarts/tree/main/managed-agents
- Agent SDK Python (adjacent library): https://github.com/anthropics/claude-agent-sdk-python
- Agent SDK TypeScript (adjacent library): https://github.com/anthropics/claude-agent-sdk-typescript
- Engineering blog: https://www.anthropic.com/engineering/managed-agents
- Trust center: https://trust.anthropic.com/
- Careers: https://www.anthropic.com/careers

## Nearby in the atlas

- [n8n](/products/n8n.md)
- [Trigger.dev](/products/trigger-dev.md)
- [Windmill](/products/windmill.md)
- [AgentNest](/products/agentnest.md)
- [Anthropic Sandbox Runtime (srt)](/products/sandbox-runtime.md)
- [Aptible](/products/aptible.md)

## Method

Facts come from the product's own public pages. Each claim carries its evidence: a quote and the page it came from. We count capabilities rather than grade products, and "Not assessed" means we have not checked, never that the product lacks it.