Docker marks Agentic Platform experimental. Stateful stop/resume depends on account enablement, and experimental volume-backed sandboxes cannot use that stop path. Read the source ↗

Agent sandboxes / Docker Cloud Sandboxes alternative

A Docker Cloud Sandboxes alternative for a growing agent computer

Docker Cloud Sandboxes can preserve memory and filesystem on supported stop/resume paths. Its independent experimental volumes have different persistence and lifecycle rules. Compare those combinations with an agent that repeatedly returns to a growing RootFS rather than treating every Docker sandbox as the same environment.

Start building Compare capabilities ↓

Reviewed 2026-10-18 · Published by the Sandbox0 team

When Docker Cloud Sandboxes fits

Docker Cloud Sandboxes fits teams that want the sbx agent workflow, kits and templates, Docker-managed cloud execution, and filesystem transfer between local and cloud sandboxes. Verify account access, supported architecture, and the documented Docker-in-sandbox limitations for the actual workflow.

When to consider Sandbox0

Consider Sandbox0 when the primary requirement is a durable complete writable RootFS that expands online and supported file operations during compute pauses. Port kits, model credentials, local integrations, and any Docker-specific development workflow separately.

Capabilities

Docker Cloud Sandboxes and Sandbox0: what persists?

Feature comparison with documentation and verification status
Capability	Docker Cloud Sandboxes	Sandbox0
Environment and runtime lifetime	Docker Cloud SandboxesCreation defaults to one hour; extensions cannot exceed 24 hours from creation. Supported stop/resume starts a new running TTL period; zero does not disable expiration.View evidence ↗	Sandbox0ttl and hard_ttl default to 0 (disabled). Configure idle pause and deletion policies for the same persistent sandbox identity; an idle running sandbox still accrues compute charges.Read documentation →
Persistent files	Docker Cloud SandboxesSupported stateful stop and reusable filesystem templates retain data; independent volumes have separate experimental exit-snapshot semantics.View evidence ↗	Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFS	Docker Cloud SandboxesStateful stop preserves the sandbox filesystem and templates capture configured disk state. Host mounts and independent volumes are not implicitly included in cross-backend moves.View evidence ↗	Sandbox0Persistence extends beyond /workspace to paths such as /usr/local, /etc, and /home. Runtime tmpfs and virtual filesystems are excluded.Read documentation →
Release compute, keep state	Docker Cloud SandboxesSupported cloud stop preserves memory/filesystem and releases compute billing; account enablement matters and volume-backed sandboxes cannot stop.View evidence ↗	Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process state	Docker Cloud SandboxesStateful stop and explicit all-state template capture include execution memory; ordinary templates and local/cloud moves are disk-only.View evidence ↗	Sandbox0Experimental memory checkpoints can preserve compatible process state. Use memory: true on both pause and resume on a compatible deployment. A filesystem checkpoint alone does not preserve running processes.Read documentation →

Online RootFS growth

On growth-enabled images, Sandbox0 automatically expands the writable RootFS as free space runs low, within a configured sparse logical ceiling up to 64 TiB. Agent processes keep running. Growth is enabled in the qualified Cloud region; confirm activation for other regions. Existing images and snapshots retain their captured storage geometry. The ceiling is logical capacity, not preallocated host disk.

The reviewed cloud guide exposes supported resource shapes, snapshots, and independent volumes; it does not establish automatic online expansion of the complete mounted RootFS as an active agent fills it.

Files while compute is paused

Sandbox0's Files API reads persistent files without resuming compute. Filesystem-only pauses also permit edits; retained-memory pauses are read-only and reject mutations with 409. SSH, SFTP, file watching, and runtime-only mounts require running compute.

See paused file operations →

Stateful stop and independent volumes cannot be freely combined

The cloud guide documents stop preserving memory and filesystem, with compute unbilled while stopped. Volume-backed sandboxes cannot use that stop path and require deletion at timeout. Experimental volume data is saved on sandbox exit rather than continuously; concurrent users can overwrite the stored volume snapshot with the last exit. Treat machine snapshots and independent volumes as different persistence choices.

Source 1 ↗

A local/cloud move is a disk copy, not process migration

Docker’s move workflow creates a separate destination from a filesystem snapshot, leaves the source in place, and does not transfer running processes or RAM. Host mounts, managed secrets, volume contents, and network policy need separate handling. Within cloud stop/resume, the sandbox identity remains, but that is a different operation from moving between backends.

Source 1 ↗Source 2 ↗

Timeouts, account quotas, and console prices shape the workload

Cloud creation defaults to a one-hour timeout; zero is invalid, and deadline extensions are bounded. A stopped resumable sandbox restarts its timer on resume. The plan has no recurring subscription fee and meters running CPU/memory configuration in seconds, with exact effective prices shown in the Console. Default account quotas are not guaranteed capacity for every account.

Source 1 ↗Source 2 ↗Source 3 ↗

Pricing

Compare billing models, then price your workload

CPU allocation, memory, billing minimums, disk, plan fees, and supported workload differ. These meters are not equal-performance machine quotes. Confirm region, resource units, and included allowances before estimating a total.

Docker Cloud Sandboxes

Pricing reviewed 2026-10-18

Cloud platform recurring fee: No recurring subscription fee; A separate personal-account Docker Agentic Platform subscription must be activated; it is not the Docker Desktop subscription.; Rate source ↗
Running cloud compute: Pay as you go, metered in seconds; Check the effective configuration quote in the Console Usage & billing page.; Rate source ↗
Supported stopped sandbox compute: $0 while stopped; Does not establish free indefinite artifact/storage retention; volume-backed sandboxes cannot use this path.; Rate source ↗
Model inference: Paid to your model provider; Rate source ↗

Do not apply local Docker’s lack of a cloud-compute meter to Docker-managed sandboxes, or assume Docker Desktop fees include this usage. Verify effective console prices, credits, stored-artifact charges, quotas, and account feature enablement. No numeric compute-savings claim is possible from the public billing description alone.

Check current Docker Cloud Sandboxes pricing ↗

Sandbox0

$0.015/GiB-hour compute (billed by the second at approximately $0.000004167/GiB-second), plus $0.02/GiB-month of retained incremental RootFS and memory checkpoint data. Shared base images and platform recovery retention are excluded; paused sandboxes accrue storage charges only.

Compute / GiB-hour: $0.015
Retained data / GiB-month: $0.02
Paused compute: $0.00

A 4 GiB agent, 100 running hours

$6.00 compute

4 GiB × 100 hours, billed by the second. Keeping 20 GiB of billable incremental data for a full billing month adds $0.40. CPU is derived from memory by the platform; choose a tier that fits your workload.

Understand storage billing →

Before migrating from Docker Cloud Sandboxes

Inventory supported stop/resume state, experimental volumes, kits, saved templates, local mounts, credentials, architectures, and services before exporting state.
Build compatible tools in a Sandbox0 template and port agent command/file, credential, network, and service access; transfer attached data independently where required.
Validate system state, growing RootFS data, paused file access, process-continuity requirements, timeout policy, and actual configuration pricing.

Common questions

Is Sandbox0 a drop-in Docker Cloud Sandboxes replacement?

Evaluate API compatibility and product scope. A shared sandbox category does not make SDKs, lifecycle semantics, or limits interchangeable.

Does keeping files also keep a running agent?

Filesystem state and volatile process state are separate. Sandbox0 requires explicit compatible memory checkpoints to preserve running processes; its default filesystem resume starts fresh processes.

Sources and review scope

The Boat team's Battleships directory is the discovery source. References below describe the product and plan discussed on this page. Check current availability, regional terms and billing units before migrating. Where the cited material does not establish a capability or price, the comparison identifies that uncertainty.

Battleships research record ↗docs.docker.com/ai/sandboxes/cloud/usage/ ↗docs.docker.com/ai/sandboxes/cloud/local-vs-cloud/ ↗docs.docker.com/subscription-billing/plans/docker-agentic-platform/ ↗docs.docker.com/ai/sandboxes-api/limits/ ↗docs.docker.com/agentic-platform/signup/ ↗Sandbox0 RootFS capacity →

Related comparisons

Agent 37 vs Sandbox0 →AgentComputer vs Sandbox0 →ainclave vs Sandbox0 →Alibaba Cloud Agent Sandbox / FC / AgentRun vs Sandbox0 →Alibaba Cloud AgentBay vs Sandbox0 →Amika vs Sandbox0 →