Agent sandboxes / Lizard alternative

A Lizard sandbox alternative beyond a node-local workspace

Lizard combines sandboxes with an application platform, databases, and storage. Its sandbox shape prices differ from its application compute rates, and its persistent volumes differ from execution snapshots. For a returning agent, trace both the saved process state and the files that remain after the machine ends.

Start buildingCompare capabilities ↓

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

When Lizard fits

Lizard fits teams that want sandbox execution beside app delivery and managed data services, with its agent templates, SDK, snapshots, and running-machine forks where supported. Runtime type depends on the template or image, so check container versus microVM behavior.

When to consider Sandbox0

Consider Sandbox0 when installed system tools and agent outputs belong in a durable whole RootFS backed by regional object storage, with automatic online growth and supported file operations during compute pauses. Validate runtime compatibility and latency with the actual workload.

Capabilities

Lizard and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeLizardCreation supports timeoutMs: 0 for no expiration; SDK and CLI defaults remain five minutes. Updating a lifetime requires a positive value.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 filesLizardSeparate /workspace volumes survive sandbox termination but are node-local and attach to one sandbox at a time.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSLizardThe reviewed snapshot contract specifies memory and /workspace; complete writable RootFS coverage outside that path is not established.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 stateLizardThe SDK exposes pause/resume with retained execution state; connect alone does not resume a paused sandbox.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateLizardSnapshots and supported running forks retain memory/processes; fork rejects attached volumes and pod-backed runtimes.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.

Lizard offers explicit persistent-volume resize with runtime-specific restrictions. Its node-local /workspace volume is a different durability and growth boundary from automatic online growth of Sandbox0’s complete persistent writable RootFS.

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 →

Persistent volumes and memory snapshots cover different state

Lizard mounts a persistent volume at /workspace, with one attached sandbox on the volume’s node. The volume remains after kill or expiration, but uses node-local storage and needs an export strategy for node failure. Snapshots save memory and workspace files, exclude an attached volume, and a volume-attached sandbox cannot be forked. Pod-backed sandboxes also reject fork. Check these combinations before designing retry fan-out.

Source 1 ↗Source 2 ↗Source 3 ↗

Explicit volume resize is separate from automatic RootFS growth

New volumes are normally 1–50 GB and can be resized while attached; Firecracker volumes only grow. This is an explicit volume operation, not evidence of automatic whole-system RootFS expansion. Without a volume, workspace files are temporary when the sandbox stops, unless retained through the appropriate supported snapshot workflow.

Source 1 ↗Source 2 ↗

Long-running settings and platform billing need their own checks

The raw API can create a sandbox with no expiration, while the SDK and CLI default to five minutes. Commands do not refresh the explicit lifetime deadline. Pro includes a monthly usage credit, and sandbox shapes have flat running rates rather than the platform’s separate application CPU/RAM meters. Paused compute and retained volume storage have separate billing treatment.

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.

Lizard

Pricing reviewed 2026-10-18

Pro subscription
$19 per month including $19 monthly usage credit
Unused monthly credit does not roll over; usage above the allowance is charged separately.
Rate source ↗
Small sandbox
$0.0162 per running hour
2 vCPU, 4 GB RAM, 5 GB workspace disk; persistent volumes separate.
Rate source ↗
Medium sandbox
$0.0324 per running hour
4 vCPU, 8 GB RAM, 20 GB workspace disk.
Rate source ↗
Large sandbox
$0.0648 per running hour
8 vCPU, 16 GB RAM, 50 GB workspace disk.
Rate source ↗
Persistent volume storage
$0.14 per GB-month
The volume guide bills used space, not the full allocation.
Rate source ↗
Sandbox outbound traffic
2 TB per account over a rolling 30 days included
No separate sandbox per-GB rate; contact the vendor for workloads beyond the allowance.
Rate source ↗

Do not apply app/database CPU, memory, or outbound traffic rates to sandbox shape pricing. Pro credit is included rather than added to all usage a second time. The seven-day trial has $5 credit and requires a card; running limits differ before the first paid invoice. Paused compute is not billed, while retained volume storage continues.

Check current Lizard 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 Lizard

  1. Export node-local volume data and classify snapshots, active memory, system changes, and any volume-attached or pod-backed fork restrictions.
  2. Rebuild needed software in the target template and port command/file, credential, service, and lifecycle calls; explicitly set the intended lifetime policy.
  3. Validate full system state across tasks, growing data, paused file access, supported process restoration, and the plan-credit plus compute/storage cost.

Common questions

Is Sandbox0 a drop-in Lizard 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 ↗lizard.build/docs/sandboxes/volumes/ ↗lizard.build/docs/sandboxes/sdk-reference/ ↗lizard.build/docs/platform/storage-and-recovery/ ↗lizard.build/docs/platform/limits/ ↗lizard.build/docs/platform/billing/ ↗lizard.build/pricing ↗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 →