Agent sandboxes / Isorun alternative

An Isorun sandbox alternative for a persistent agent computer

Isorun offers both saved runtime state and an independent persistent disk mounted at /workspace. Those are different retention paths: a workspace disk can outlive a destroyed VM, while a runtime checkpoint captures the prepared execution environment.

Start buildingCompare capabilities ↓

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

When Isorun fits

Isorun fits teams using microVM execution, OCI images, hibernation, checkpoint branching, and Tigris-backed workspace disks. Independent copy-on-write disks can separate durable datasets from disposable compute.

When to consider Sandbox0

Consider Sandbox0 when system packages, configuration, and workspace outputs should share one durable writable RootFS that grows automatically. Supported paused Files API operations can manage persistent state without resuming the agent computer.

Capabilities

Isorun and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeIsorunNot established in this reviewSandbox0ttl 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 filesIsorunAn independent disk mounted at /workspace survives sandbox destruction and can attach to another sandbox.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSIsorunA saved runtime checkpoint captures filesystem state; retaining only the independent workspace disk has a narrower scope.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 stateIsorunHibernation keeps the runtime ID and saved state for resume.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateIsorunThe checkpoint guide includes loaded runtime state, while outbound sessions and in-flight operations reset.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.

An Isorun persistent disk has a provisioned size and returns ENOSPC when full, despite billing only stored data. Scratch-disk growth and snapshot reuse do not establish automatic expansion of a complete durable 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 →

A workspace disk and a runtime checkpoint retain different state

Isorun’s independent disk is mounted at /workspace and survives sandbox destruction. Its runtime checkpoint captures filesystem and loaded runtime state, and can seed restored or forked sandboxes. Retain the appropriate artifact for packages outside the workspace and reconnect network clients after restoration.

Source 1 ↗Source 2 ↗

Sparse usage billing is different from automatic capacity growth

The persistent-storage guide bills actual data rather than the provisioned size, but says a full provisioned disk returns ENOSPC. Its lifecycle guide also describes a small scratch disk that grows on write. Neither statement should be used as proof of automatic online expansion of the entire durable root filesystem.

Source 1 ↗Source 2 ↗

Zero hibernated runtime cost still leaves independent storage

The compute price page stops active runtime billing during hibernation. Independent persistent disks have their own stored-data fee and survive a destroyed sandbox. Include those retained disks and snapshots in the budget instead of treating no active runtime charge as no retained-resource cost.

Source 1 ↗Source 2 ↗

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.

Isorun

Pricing reviewed 2026-10-18

Running CPU
$0.025 per vCPU-hour
Rate source ↗
Running RAM
$0.015 per GiB-hour
Rate source ↗
Independent persistent disk
$0.02 per GB-month of stored data
Actual stored data, not the provisioned size; retained independently of sandbox compute.
Rate source ↗

Compute is metered per second with no minimum and no active-runtime charge while hibernated. The storage guide has a separate persistent-disk meter and no egress fee. It describes clean detach flushing object-backed writes and possible loss of an unflushed tail on host failure; verify durability and snapshot retention requirements for the workload.

Check current Isorun 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 Isorun

  1. Inventory runtime snapshots, hibernated state, independent disk names, provisioned capacities, and files outside /workspace.
  2. Export the needed system and workspace state, port OCI images, tools, credentials, and network integration, and reconnect external sessions explicitly.
  3. Validate repeated tasks, installed-package retention, automatic RootFS growth, paused file access, and retained-disk costs after compute stops.

Common questions

Is Sandbox0 a drop-in Isorun 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.isorun.ai/sandboxes/persistent-storage ↗docs.isorun.ai/sandboxes/checkpoints-and-rollback ↗docs.isorun.ai/sandboxes/lifecycle ↗docs.isorun.ai/getting-started/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 →