Agent sandboxes / StateSet Sandbox alternative

A StateSet Sandbox alternative for long-lived system state

StateSet combines Kubernetes sandbox execution with file checkpoints, artifacts, and agent sessions that can rotate onto fresh sandboxes. Shared and dedicated capacity have substantially different prices and interruption behavior. Compare the actual runtime class and state being restored before choosing a long-running agent environment.

Start buildingCompare capabilities ↓

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

When StateSet Sandbox fits

StateSet fits teams using its managed agent sessions, checkpoint/restore workflow, artifact storage, usage tracking, and session budgets. Shared capacity can suit retryable batch work; its dedicated class is advertised for work that needs guaranteed resources without Spot preemption.

When to consider Sandbox0

Consider Sandbox0 when the same complete durable writable RootFS should retain evolving system tools and expand online across tasks, with supported file operations during compute pauses. Preserve or supply the surrounding agent-session orchestration separately.

Capabilities

StateSet Sandbox and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeStateSet SandboxPro and Team advertise up to 24-hour sandbox runtime with separate three-/seven-day durable lifetime and session-rotation behavior.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 filesStateSet SandboxFile checkpoints can restore into the same or a new sandbox; durable capacity and retention depend on plan.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSStateSet SandboxPath-filterable filesystem checkpoints do not establish uniform complete operating-system retention for every runtime/lifecycle path.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 stateStateSet SandboxPaid durable sandboxes support compute-free pause; agent-session pause is a different state with its sandbox still alive.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateStateSet SandboxThe reviewed checkpoint contract specifies files, environment and metadata, not arbitrary process-memory restoration.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.

StateSet’s documented checkpoints, artifacts, and session rotation preserve useful application work, but the reviewed guides do not establish automatic online growth of an entire durable RootFS while the existing agent keeps running.

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 →

A rotated agent session is not the same running process

The agent-session guide checkpoints a nearing-expiry sandbox, provisions a fresh one, and restores files/context. The checkpoint guide describes filesystem, environment variables, and metadata with path filters. Those artifacts can continue application progress, but do not establish arbitrary live process-memory continuation or uniform complete-RootFS coverage.

Source 1 ↗Source 2 ↗

Session pause and compute-free durable pause have different meanings

The sessions guide says a paused session keeps its sandbox alive while stopping the rotation timer. The price page separately says paused durable sandboxes stop compute billing. Plan limits also differ: Hobby has no durable sandboxes; Pro and Team list three- and seven-day durable lifetimes. Match the actual operation and retained state instead of assuming every pause stops billing.

Source 1 ↗Source 2 ↗

Dedicated billing uses half the allocation, not sampled busy time

Shared CPU/RAM is billed on full provisioned size and may be preempted. Dedicated list rates are applied to 50% of allocated CPU and RAM, described as an average-utilization estimate; this is not actual-consumption metering. Included compute allowances apply only to dedicated usage, and plan fees, creation events, and egress remain separate.

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.

StateSet Sandbox

Pricing reviewed 2026-10-18

Hobby / Pro / Team platform fees
$0 / $49 / $299 per month, plus usage
Rate source ↗
Shared CPU
$0.0055 per vCPU-hour
Full size, wall-clock running time; preemptible capacity.
Rate source ↗
Shared RAM
$0.00075 per GB-hour
Rate source ↗
Shared 4 vCPU / 8 GB example
$0.028 per hour
No included shared-compute allowance.
Rate source ↗
Dedicated Hobby / Pro list rates
$0.128 per vCPU-hour and $0.0106 per GB-hour
Charged on 50% of allocation; effective 4 vCPU/8 GB example is $0.2984/hour.
Rate source ↗
Dedicated Team list rates
$0.10 per vCPU-hour and $0.008 per GB-hour
Charged on 50% of allocation; 4 vCPU/8 GB example is $0.232/hour.
Rate source ↗
Egress Hobby / Pro / Team
$0.15 / $0.15 / $0.12 per GB
Rate source ↗
Creation events Hobby / Pro / Team
$0.60 / $0.60 / $0.50 per million
Rate source ↗

Shared and dedicated prices do not represent the same resource guarantee. Monthly included CPU/RAM applies only to dedicated usage; shared bills from the first second. Egress/creation allowances, one-time trial credits, retained data and effective plan limits require separate accounting. The 50% dedicated factor is a fixed allocation rule, not a promise of CPU/RAM usage sampling or equivalent throughput to Sandbox0.

Check current StateSet Sandbox 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 StateSet Sandbox

  1. Inventory shared versus dedicated class, plan/retention limits, checkpoint include/exclude paths, session context, artifacts, budgets, and rotation behavior.
  2. Export required files/context and port compatible templates, execution, secrets, networking, and the surrounding agent-session orchestration.
  3. Validate system-package retention, online disk growth, paused file operations, process/restart expectations, checkpoint recovery and full plan/usage charges.

Common questions

Is Sandbox0 a drop-in StateSet Sandbox 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 ↗sandbox.stateset.app/docs/agent-sessions ↗sandbox.stateset.app/docs/checkpoints ↗sandbox.stateset.app/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 →