# Docker Sandboxes

Run agents safely, local to cloud. — Run Claude Code, Codex, and your other agents at full speed in MicroVM environments. Local or cloud. One CLI.

Source: https://wheredoirunthis.com/products/docker-sandboxes · JSON: https://wheredoirunthis.com/api/products/docker-sandboxes · Read on 2026-10-08

## At a glance

- Type: Sandbox
- Kind: Hosted
- Isolation: Virtual machine
- Free tier: Yes
- Starting price: $0.07/h
- GPU: No
- Regions: Not disclosed
- Coverage: 22 of 61 assessed capabilities supported
- Agent readiness: read 11 of 11 signals, drive 10 of 11 signals

## Links

- Homepage: https://www.docker.com/products/docker-sandboxes/
- Docs: https://docs.docker.com/ai/sandboxes/
- Pricing: https://www.docker.com/products/docker-sandboxes/
- Repository: https://github.com/docker/sbx-releases

## Pricing

- Model: Local sandboxes: free (incl. commercial). Cloud sandboxes: pay-as-you-go, metered per second by vCPU+memory size, no recurring subscription fee; one-time promotional compute credit at signup; BYO model-provider keys (inference billed by your provider). Org governance: separate paid subscription (sales).
- Unit: sandbox-second (by size: vCPU + memory); shown as an hourly rate
- Free tier: Yes

| Tier | Price | Includes | Notes |
| --- | --- | --- | --- |
| Local | Free | sbx CLI + local microVM sandboxes on your own Mac/Windows/Linux machine, incl. commercial use. Docker account sign-in required. |  [source](https://docs.docker.com/ai/sandboxes/faq/) |
| Cloud — Micro (1 vCPU / 2 GB) | $0.07/hour, metered per second | Docker-managed cloud sandbox, same microVM model |  [source](https://www.docker.com/products/docker-sandboxes/) |
| Cloud — Small (2 vCPU / 4 GB) | $0.14/hour, metered per second | Default size for CLI + SDK kit launches |  [source](https://www.docker.com/products/docker-sandboxes/) |
| Cloud — Medium (4 vCPU / 8 GB) | $0.28/hour, metered per second |  |  [source](https://www.docker.com/products/docker-sandboxes/) |
| Cloud — Large (8 vCPU / 16 GB) | $0.56/hour, metered per second |  |  [source](https://www.docker.com/products/docker-sandboxes/) |
| Cloud — XL (16 vCPU / 32 GB) | $1.12/hour, metered per second |  | Linear: $0.07 per vCPU-hour (with 2 GB RAM per vCPU) across all sizes. [source](https://www.docker.com/products/docker-sandboxes/) |
| AI Governance (org policy) | Contact sales | Org-wide network, filesystem and MCP policy for local sandboxes, audit/SIEM export | 'Available on a separate paid subscription' — price not public. [source](https://www.docker.com/products/ai-governance/) |

- Overages: Pure metering — no included quota to exceed beyond the one-time signup credit (amount shown only in the Console). Volume / snapshot storage pricing not disclosed publicly. Compute not billed while a sandbox is stopped.
- Card required: Yes
- Source: https://docs.docker.com/agentic-platform/signup/

## Runtime

- **GPU & accelerators**: No [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/)
- **BYOC (Bring your own cloud)**: no — "Cloud runs only on 'Docker-managed infrastructure'. No BYOC. The functional alternative is the free LOCAL mode on your own machine (not BYOC — the developer's laptop/workstation, not a cloud account)." [source](https://docs.docker.com/ai/sandboxes/cloud/) (confidence: medium)
- **Secrets management**: Yes [source](https://docs.docker.com/ai/sandboxes/architecture/)
- **Storage options**: Sandbox filesystem persists across supported stops; cloud stop preserves memory and filesystem. Local workspaces use host filesystem passthrough; cloud has no host mounts and uses file transfer or in-sandbox clone. Experimental cloud volumes persist independently, save snapshots only on exit, and use last writer wins. Volume-backed sandboxes cannot stop; API cloud volume attachment is account-gated, one exclusive entry, and unsupported for memory-snapshot images. [source](https://docs.docker.com/ai/sandboxes/cloud/usage/)
- **Supported workloads**: AI coding agents in isolated sandboxes (cloud SDK catalog: Claude Code, Codex, Cursor, Devin, Gemini CLI, OpenCode; Console also advertises Hermes; other dedicated agent pages describe local sandboxes), agents requiring a private Docker Engine, subject to documented cloud docker exec/compose exec/healthcheck filesystem limitations, programmatic code execution via REST API and JavaScript/TypeScript SDK, detached cloud agent tasks, on-demand HTTP apps behind public HTTPS ports with account-supported auto-resume [source](https://www.docker.com/products/docker-sandboxes/)
- **Persistence**: both — "Local persists across stops until removed. Cloud defaults to a 1h TTL; expiry stops resumable sandboxes and deletes others. TTL pauses while stopped and starts a new period on resume; extensions are limited to 24h from creation. Stop/resume may be disabled by account. Volume-backed sandboxes cannot stop and require delete on timeout. Snapshots, including pause snapshots, expire after seven days of non-use. Always-on restart requires account access." [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/) (confidence: high)
- **Data jurisdiction**: United States — explicitly disclosed for Cloud Sandboxes; specific cloud regions not disclosed — "Docker’s self-service AI services (including the Docker Agentic Platform, cloud sandboxes, and MCP Gateway) are hosted on infrastructure located in the United States." [source](https://www.docker.com/legal/privacy/) (confidence: high)
- **Resource limits & quotas**: Cloud sizes micro (1 vCPU/2 GiB) to xl (16 vCPU/32 GiB). Default quotas: 10 concurrent, 50 stored sandboxes, 100 volumes, 100 secrets, 3 images preparing. Default TTL 1h; extensions cannot exceed 24h from creation, but TTL pauses while stopped and starts a new period on resume. Stop/volume/always-on availability is account-dependent. Rate limits not disclosed numerically. Console is individual-use only. [source](https://docs.docker.com/ai/sandboxes-api/limits/)
- **SDK languages**: TypeScript [source](https://docs.docker.com/ai/sandboxes-api/install/)
- **Network & egress controls**: Yes [source](https://docs.docker.com/ai/sandboxes/cloud/network-policy/)
- **Snapshots & forks**: Yes [source](https://docs.docker.com/ai/sandboxes-api/cookbook/snapshot-and-fork-a-sandbox/)
- **Deployment methods**: cli, api, sdk, dashboard, container-image, template — "`sbx` CLI (`sbx run claude`, `sbx --cloud create`); REST API at https://connect.docker.com/sandboxes; TypeScript SDK; Docker Agentic Platform web Console; create from a registry image (`imageRef`) or a registered image; launch from a Kit (bundled catalog / OCI kit reference)." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) (confidence: high)
- **Isolation**: Virtual machine — "'No, these aren't containers. … microVM environments we built from scratch. Each one gets its own kernel, its own private Docker Engine' + 'A VMM we built ourselves' (landing); docs: 'Every sandbox runs inside a lightweight microVM with its own Linux kernel.' Hypervisor is Docker's OWN VMM (not Firecracker/Kata) → `vm` (own-kernel, hypervisor not a named enum value). Local needs KVM (Linux) / Windows Hypervisor Platform. CLOUD substrate is asserted by marketing as 'Same microVM boundary on both sides, no downgrade' but the cloud hypervisor/provider is not otherwise documented (medium confidence for the cloud side)." [source](https://www.docker.com/products/docker-sandboxes/) (confidence: high)

## Capabilities

### AI & agents

| Capability | Status | Evidence |
| --- | --- | --- |
| Agent skills | Supported | "Shared agent skills let you install skills from Git repositories or import skills from supported agents on your host" [source](https://docs.docker.com/ai/sandboxes/workflows/agent-skills/) |
| AI app builder | Not supported | "No prompt-to-app surface — sandboxes run third-party coding agents." [source](https://www.docker.com/products/docker-sandboxes/) |
| Autonomous QA agent | Not supported | "No first-party QA agent." [source](https://docs.docker.com/ai/sandboxes/) |
| Built-in coding agent | Supported | "Bundled kits: `docker-agent` \| Docker Agent" [source](https://docs.docker.com/ai/sandboxes/agents/docker-agent/) |
| BYOK for LLMs | Supported | "Bring your own key for model inference." [source](https://www.docker.com/products/docker-sandboxes/) |
| Credential-injection proxy | Supported | "The proxy rewrites the header: The proxy replaces the sentinel with the real credential after the request has left the microVM." [source](https://docs.docker.com/ai/sandboxes/architecture/) |
| LLM gateway | Not supported | "The proxy replaces the sentinel with the real credential after the request has left the microVM." [source](https://docs.docker.com/ai/sandboxes/architecture/) |
| MCP client | Supported | "The gateway gives the agent inside the sandbox one MCP endpoint, while sbx manages the registered servers, OAuth credentials, and sandbox lifecycle on the host." [source](https://docs.docker.com/ai/sandboxes/mcp-gateway/) |
| MCP-server hosting | Partially supported | "For --local --url registrations that resolve to OCI packages, use a host with Docker installed" [source](https://docs.docker.com/ai/sandboxes/mcp-gateway/) |
| MCP server | Partially supported | "MCP endpoint for structured agent access: https://mcp-docs.docker.com/mcp" [source](https://docs.docker.com/llms.txt) |
| Model inference API (first-party) | Not supported | "Bring your own key for model inference." [source](https://www.docker.com/products/docker-sandboxes/) |
| Preinstalled coding agent | Supported | "The SDK bundles a catalog of kits you can launch by name." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |

### Networking & delivery

| Capability | Status | Evidence |
| --- | --- | --- |
| Automatic HTTPS | Supported | "The command returns a public HTTPS URL assigned by the cloud control plane." [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Custom domains | Not supported | "Use the URL returned by the service rather than constructing a hostname." [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Direct public IPv6 | Not supported | "Host IP addresses, host port bindings, and other protocols are rejected." [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Edge KV store | Not applicable | "n/a" [source](https://docs.docker.com/ai/sandboxes/) |
| Egress controls | Supported | "All HTTP/HTTPS egress routes through a host-side proxy enforcing a domain allowlist; raw TCP, UDP, ICMP … blocked" [source](https://docs.docker.com/ai/sandboxes/security/isolation/) |
| Email send / receive | Not supported | "No email primitive documented." [source](https://docs.docker.com/ai/sandboxes/) |
| Public preview URLs | Supported | "$ sbx --cloud ports cloud-project --publish 8080 — returns a public HTTPS URL" [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Private networking | Not supported | "each sandbox has its own isolated network; sandboxes cannot reach each other or the host's localhost" [source](https://docs.docker.com/ai/sandboxes/security/isolation/) |
| Built-in reverse proxy | Supported | "Reach a web application running inside a sandbox through an HTTPS URL. Publishing creates a route to the application" [source](https://docs.docker.com/ai/sandboxes-api/cookbook/expose-a-port-from-a-cloud-sandbox/) |

### Organisation & security

| Capability | Status | Evidence |
| --- | --- | --- |
| Built-in auth | Not supported | "Docker account authentication does not protect the application URL. Configure authentication in your application" [source](https://docs.docker.com/ai/sandboxes-api/cookbook/expose-a-port-from-a-cloud-sandbox/) |
| BYOC | Not supported | "Cloud sandboxes run AI agents on Docker-managed infrastructure instead of your local machine." [source](https://docs.docker.com/ai/sandboxes/cloud/) |
| Deploy into customers' VPCs | Not supported | "A sandbox kit defines an environment for an agent or tool, including its image, setup, network rules, and credential requirements." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |
| Secrets management | Supported | "Secrets provide credentials without placing their values inside a sandbox." [source](https://docs.docker.com/ai/sandboxes-api/cookbook/manage-cloud-secrets/) |
| Self-host / OSS core | Partially supported | "The sbx CLI and local sandbox compute are free to use, including for commercial work." [source](https://docs.docker.com/ai/sandboxes/faq/) |
| Sharing | Partially supported | "Treat an exposed URL as a public endpoint." [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| SSO / SAML | Partially supported | "Browser sign-in supports single sign-on and two-factor authentication." [source](https://docs.docker.com/ai/sandboxes-api/authentication/) |
| Teams | Partially supported | "The initial release is for individual use… Shared workspaces and team ownership aren't supported." [source](https://docs.docker.com/agentic-platform/faq/) |
| Workload identity | Not supported | "A sandbox kit defines an environment for an agent or tool, including its image, setup, network rules, and credential requirements." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |

### Developer access

| Capability | Status | Evidence |
| --- | --- | --- |
| CLI | Supported | "$ brew trust docker/tap && brew install docker/tap/sbx" [source](https://www.docker.com/products/docker-sandboxes/) |
| E2B-compatible API | Not supported | "A sandbox kit defines an environment for an agent or tool, including its image, setup, network rules, and credential requirements." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |
| Filesystem API | Supported | "GET /v1/files/download … readFile, writeFile, upload, mkdir, move, stat, remove" [source](https://docs.docker.com/reference/api/sandboxes/latest/) |
| Git deploy | Not supported | "Git branch mode gives each agent its own worktree" [source](https://docs.docker.com/ai/sandboxes/workflows/git/) |
| HTTP data API | Not applicable | "n/a" [source](https://docs.docker.com/ai/sandboxes/) |
| OpenAPI spec | Supported | "[OpenAPI specification](/reference/api/sandboxes/api.yaml)" [source](https://docs.docker.com/reference/api/sandboxes/latest/) |
| Private container registry | Partially supported | "Registration creates an image record and a temporary upload destination. You must push the image content separately with an OCI-compatible registry client." [source](https://docs.docker.com/ai/sandboxes-api/cookbook/register-and-manage-an-image/) |
| Public API | Supported | "Use the Docker Sandboxes API to create cloud sandboxes, run commands, and transfer files" [source](https://docs.docker.com/ai/sandboxes-api/) |
| SDK | Supported | "$ npm install @docker/sandboxes" [source](https://docs.docker.com/ai/sandboxes-api/install/) |
| SSH access | Supported | "$ sbx --cloud setup ssh / $ ssh sbx_01abc123@sbx_cloud" [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |

### Compute & storage

| Capability | Status | Evidence |
| --- | --- | --- |
| Cloud-init / ignition config | Not supported | "Kits: declarative YAML applied at startup" [source](https://docs.docker.com/ai/sandboxes/customize/use-kits/) |
| Desktop / computer-use sandbox | Not supported | "GPU, USB, display, and nested virtualization options [local only]" [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/) |
| GPU compute | Not supported | "Hardware: Has no access to host hardware" [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/) |
| Live-state (memory) snapshots | Supported | "Stop a cloud sandbox while preserving its memory and filesystem:" [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Multi-region | Not supported | "CPU architecture: Supports linux/amd64 and linux/arm64, subject to account availability" [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/) |
| Object storage | Not supported | "Cloud volumes are experimental and preserve data independently of a sandbox." [source](https://docs.docker.com/ai/sandboxes/cloud/usage/) |
| Persistent disk | Supported | "Storage: Persists in the cloud sandbox and optional cloud volumes" [source](https://docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/) |
| Sandbox templates / custom images | Supported | "Registry image (imageRef) \| A container image to use with your own sandbox settings" [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |
| Scale to zero | Supported | "Let a request to a published application start a stopped sandbox." [source](https://docs.docker.com/ai/sandboxes-api/cookbook/let-a-stopped-sandbox-resume-on-demand/) |
| Selectable isolation runtime | Not supported | "A sandbox kit defines an environment for an agent or tool, including its image, setup, network rules, and credential requirements." [source](https://docs.docker.com/ai/sandboxes-api/concepts/) |
| Snapshots / forking | Supported | "Save a sandbox's state so you can start another sandbox from the same point later… restoring it creates a new sandbox with its own identity." [source](https://docs.docker.com/ai/sandboxes-api/cookbook/snapshot-and-fork-a-sandbox/) |

### Data

| Capability | Status | Evidence |
| --- | --- | --- |
| Database branching | Not applicable | "n/a" [source](https://docs.docker.com/ai/sandboxes/) |
| Ephemeral instant DB (no-signup) | Not applicable | "n/a" [source](https://docs.docker.com/ai/sandboxes/) |
| First-party database | Not supported | "No managed database offered." [source](https://docs.docker.com/ai/sandboxes-api/) |
| Vector search | Not applicable | "n/a" [source](https://docs.docker.com/ai/sandboxes/) |

### Platform & operations

| Capability | Status | Evidence |
| --- | --- | --- |
| Managed CI/CD runners | Not supported | "No CI runner product (Docker Build Cloud is a separate sibling)." [source](https://docs.docker.com/ai/sandboxes/) |
| Built-in observability | Partially supported | "listPolicyLogEntries … PolicyLogEntry" [source](https://docs.docker.com/reference/api/sandboxes/latest/) |
| Preview environments | Not supported | "No per-PR ephemeral environment feature." [source](https://docs.docker.com/ai/sandboxes/) |
| Scheduled tasks | Not supported | "No cron/scheduler primitive documented." [source](https://docs.docker.com/ai/sandboxes/workflows/automation/) |
| Project template marketplace | Not applicable | "Browse Kits for curated and community agents" [source](https://docs.docker.com/agentic-platform/faq/) |
| Durable workflows | Not supported | "No durable-execution primitive." [source](https://docs.docker.com/ai/sandboxes-api/) |

## Agent readiness

### Read: 11 of 11 signals present

11/11 signals (7 full, 4 partial)

- Llms txt: pass — "docs.docker.com/llms.txt links Sandboxes, API/SDK, API reference, Agentic Platform, Docker Skills, page Markdown routes and docs MCP. www.docker.com/llms.txt returns 404."
- Llms full txt: partial — "docs.docker.com/llms-full.txt (465,290 bytes) remains a page-metadata and Markdown-URL index, not full documentation text."
- Robots AI policy: pass — "docs.docker.com/robots.txt: 'Content-Signal: ai-train=yes, search=yes, ai-input=yes' — explicit AI allow. (www robots.txt is silent default-allow.)"
- JSON LD: partial — "Product page JSON-LD: Corporation + FAQPage graphs; no SoftwareApplication/Product with offers."
- Sitemap: pass — "www robots → sitemap_index.xml (post/page sitemaps); docs.docker.com/sitemap.xml 372 KB incl. every sandbox + API page."
- Meta quality: pass — "<title>Docker Sandboxes | Sandboxes for Coding Agents | Docker</title> + description + og:title/description/image + twitter:card."
- Markdown negotiation: pass — "docs.docker.com returns text/markdown for 'Accept: text/markdown' and serves <page>.md for every page; www marketing page does not."
- Content efficiency: partial — "Fresh landing HTML: 503,660 bytes; readable text / HTML = 2.10% (stdlib HTMLParser excluding script/style/noscript)."
- Docs discoverable: pass — "Landing links 'Read the docs' (docs.docker.com/ai/sandboxes/) in the hero; plain HTML links, no JS needed."
- Pricing machine readable: pass — "Landing 'Pay only for what you use' HTML table: Size / vCPUs / GB / Per hour for Micro→XL ($0.07→$1.12)."
- Anti bot: partial — "Marketing and docs fetchable; public hiring embed and posting API blocked by HTTP 403. Well-known MCP descriptor also returned 403."

### Drive: 10 of 11 signals present

10/11 signals (5 full, 5 partial)

- Public API: pass — "Documented Sandboxes REST API (management at connect.docker.com/sandboxes + per-sandbox endpoint) — experimental."
- Openapi spec: pass — "Fetchable OpenAPI 3.2.0 at docs.docker.com/reference/api/sandboxes/api.yaml (58 operations)."
- MCP server: partial — "Only a docs-search MCP (https://mcp-docs.docker.com/mcp, advertised in llms.txt); no MCP server for operating sandboxes."
- SDK packages: partial — "One official SDK: TypeScript (@docker/sandboxes)."
- Auth docs: pass — "Authentication doc: OAuth device/browser sign-in, PAT with `sandbox:use` scope, PAT→access-token exchange, per-sandbox scoped endpoint tokens, permission names."
- Agent self serve creds: partial — "Human signup + card checkout to activate the subscription, then a PAT created in the Docker account UI; token exchange thereafter is programmatic."
- Webhooks: fail — "No webhooks documented."
- CLI tool: pass — "Official `sbx` CLI (brew/winget/apt)."
- Error model: pass — "Errors doc + OpenAPI Error/ErrorInfo/QuotaFailure/RetryInfo schemas: 'Errors contain a stable code, a message, and optional details.'"
- Rate limit docs: partial — "Rate limits and quotas documented conceptually (default quotas table; 'honor any delay specified by the server'), but no numeric request-rate limits or rate-limit headers."
- Agent config: partial — "Docker Skills is now an official documented SKILL.md collection (https://docs.docker.com/ai/skills/); coverage is product-dependent, not a promise of a Sandboxes-specific skill. One skills-listing surface; llms.txt and docs MCP additionally aid discovery."

## Coding harnesses

- Claude Code (Anthropic) — "$ sbx run claude ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/claude-code/)
- Codex CLI (OpenAI) — "$ sbx run codex ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/codex/)
- Cursor (Anysphere) — "$ sbx run cursor ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/cursor/)
- GitHub Copilot (GitHub) — "$ sbx run copilot ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/copilot/)
- Gemini CLI (Google) — "Bundled kits: … `gemini` | Gemini CLI" [source](https://docs.docker.com/ai/sandboxes-api/concepts/)
- OpenCode (opencode) — "$ sbx run opencode ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/opencode/)
- Hermes Agent (Nous Research) — "You can run agents such as Claude Code, Codex, and Hermes in isolated cloud sandboxes with live terminals. (kit: docker.io/sbx/hermes-agent-kit)" [source](https://docs.docker.com/agentic-platform/faq/)
- Kiro CLI (Amazon Web Services) — "$ sbx run kiro ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/kiro/)
- Droid (Factory) — "$ sbx run droid ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/droid/)
- Devin (Cognition) — "$ sbx run devin ~/my-project" [source](https://docs.docker.com/ai/sandboxes/agents/devin/)

Plausible compatibility (derived, not stored):

- Aider
- Amp
- Antigravity
- Goose
- Kilo Code
- Pi
- Qwen Code

## Compliance & regions

- GDPR: self-attested (scope: hosted) — "Docker complies with leading privacy regulations like GDPR, CCPA, CPA, CTDPA, VCDPA, UCPA, and The APEC Privacy Framework." [source](https://www.docker.com/trust/compliance/)
- CCPA: self-attested (scope: hosted) — "Docker complies with leading privacy regulations like GDPR, CCPA, …" [source](https://www.docker.com/trust/compliance/)
- Regions note: US infrastructure is explicitly disclosed for Cloud Sandboxes. Named cloud regions and region selection remain undisclosed; do not infer AWS North Virginia from the company-wide subprocessor list. Local mode runs on the developer’s machine.

## Company & presence

- Headquarters: Palo Alto, CA, USA (mailing address); fully remote company with office hubs in Paris and Seattle
- Founded: 2013 (Docker first demoed publicly at PyCon, 2013-03-15, by founder Solomon Hykes; the company began as dotCloud)
- Funding: Series C, $105M, 2022-03-31
- Privacy policy: https://www.docker.com/legal/privacy/
- Terms of service: https://www.docker.com/legal/docker-subscription-service-agreement/
- GitHub: https://github.com/docker/sbx-releases
- Blog: https://www.docker.com/blog/
- X/Twitter: https://x.com/docker
- LinkedIn: https://www.linkedin.com/company/docker
- YouTube: https://www.youtube.com/user/dockerrun

## Nearby in the atlas

- [Agent Safehouse](/products/agent-safehouse.md)
- [Agent Sandbox](/products/agent-sandbox.md)
- [Agent Substrate](/products/agent-substrate.md)
- [Agent37](/products/agent37.md)
- [AgentNest](/products/agentnest.md)
- [Anthropic Sandbox Runtime (srt)](/products/sandbox-runtime.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.