Agent sandboxes / machine0 alternative

A machine0 alternative for growing agents

machine0 offers long-running cloud VMs, prepared Ubuntu/NixOS images, agent profiles, snapshots and separate disks. It can retain the whole boot filesystem; the key questions are how suspension restores the machine, what remains billed and whether accumulated system state can grow online.

Start building Compare capabilities ↓

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

When machine0 fits

machine0 fits teams wanting full VMs, Docker/kernel compatibility, CLI/MCP management, agent profiles and optional GPU capacity. Its image and provider choices serve a different compatibility envelope from a gVisor sandbox, so test native requirements before choosing a replacement.

When to consider Sandbox0

Consider Sandbox0 for compatible agents whose installed tools, configuration and outputs should remain on a durable RootFS that grows online, with supported file operations while guest compute stays paused. Use an explicit compatible experimental memory checkpoint when retaining processes is required.

Capabilities

machine0 and Sandbox0: what persists?

Feature comparison with documentation and verification status
Capability	machine0	Sandbox0
Environment and runtime lifetime	machine0Machines can remain running; credit exhaustion triggers snapshot/destruction policy rather than an unlimited prepaid entitlement.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	machine0Suspend retains a boot image, while independent beta disks have a separate lifecycle and bill.View evidence ↗	Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFS	machine0VM images retain system software and files, including on-disk profile/credential files.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	machine0Suspend saves an image and deletes compute; resume recreates a machine with a new IP.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	machine0The reviewed suspend/image documentation establishes disk-image recovery, not guaranteed retention of arbitrary live process memory.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.

machine0’s price table associates boot-disk capacity with the selected VM size; larger saved images need a destination size with enough disk. Independent disks can be 10–16,384 GB but are a separate mounted product. These reviewed contracts do not establish automatic online expansion of the complete writable boot RootFS as an 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.

The reviewed machine/image guides do not establish an API for editing a suspended boot image without recreating compute. Sandbox0 supports paused RootFS reads and filesystem-only paused mutations; memory-paused writes are rejected to preserve checkpoint state.

See paused file operations →

Stopped machines still bill; suspension deletes the cloud instance

machine0 stop keeps the instance and reserved resources and bills the full machine rate. Suspend creates an image, deletes the cloud instance and retains image-storage charges; independent disks continue billing. Resume recreates compute and gets a new IP. Deleting the suspend image also destroys that suspended machine, because it is the only saved boot-disk copy.

Source 1 ↗

Full boot images and independent disks have different ownership

Custom images capture installed system software and files and can create new VMs. Separate beta disks outlive VM destruction, are regional and single-attach, and are explicitly not backed up. They currently exclude GPU and NixOS VMs. Account for these placement and backup rules instead of assuming every mounted file is protected by a VM image.

Source 1 ↗Source 2 ↗

Snapshotting also retains on-disk agent credentials

The image guide warns that profile environment files and injected agent/provider credentials are copied into snapshots. Scrubbing files at first boot does not remove their bytes from the saved image. Review credentials before cloning or sharing an image and rotate secrets when the saved copy should no longer authorize access.

Source 1 ↗

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.

machine0

Pricing reviewed 2026-10-18

Small CPU VM: $0.013/hour; 1 vCPU, 1 GB RAM and 25 GB boot disk; billed per minute.; Rate source ↗
Large CPU VM: $0.052/hour; 2 vCPUs, 4 GB RAM and 80 GB boot disk.; Rate source ↗
XL CPU VM: $0.104/hour; 4 vCPUs, 8 GB RAM and 160 GB boot disk.; Rate source ↗
Images and suspended boot state: Approximately $0.078 per GB-month; Accrues hourly; stop still bills the full machine rate.; Rate source ↗
Independent beta disks: Approximately $0.1667 per GB-month; Billed hourly while READY, attached or not, including VM suspension.; Rate source ↗
Bandwidth overage: $0.01 per GiB; A size-dependent 1–10 TB allowance applies first.; Rate source ↗

Rates are resource bundles, not standalone vCPU or RAM prices, and GB disk/RAM labels are preserved. Standard, NVMe and premium CPU classes differ; premium explicitly uses dedicated vCPUs. Add retained images, independent disks and relevant network use. A failed resume can wait for provider capacity; no instant-wake guarantee follows from the rate table.

Check current machine0 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 machine0

Create and verify required boot images and export important files, recording independent disk regions, attachments and backup status before destroying a source.
Recreate compatible tools and credentials in Sandbox0, testing Docker/kernel/GPU dependencies separately and replacing static-IP assumptions where needed.
Validate repeated lifecycle returns, whole-RootFS growth, explicit process requirements, paused file access and full compute/image/disk/network charges.

Common questions

Is Sandbox0 a drop-in machine0 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.machine0.io/platform/vm-management.md ↗docs.machine0.io/platform/images.md ↗docs.machine0.io/platform/disks.md ↗docs.machine0.io/introduction/faq.md ↗docs.machine0.io/introduction/pricing.md ↗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 →