The public roadmap distinguishes the current single-node release from planned hibernation, fork, and user snapshots. Managed access and prices are arranged per team. Read the source ↗

Agent sandboxes / OmniRun alternative

An OmniRun sandbox alternative for a returning agent computer

OmniRun combines Firecracker execution, agent templates, desktop tools, and a choice of vendor-operated or self-operated infrastructure. Its fast template boot is different from retaining an agent’s own changed machine between tasks. Review the current lifecycle before choosing where long-lived state belongs.

Start buildingCompare capabilities ↓

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

When OmniRun fits

OmniRun fits teams that want its desktop and managed-agent integrations, Finland-hosted managed service by agreement, or the ability to operate its Apache-2.0 software on their own KVM-capable Linux hosts. Check current source availability and the exact release before planning a self-operated deployment.

When to consider Sandbox0

Consider Sandbox0 when the product needs a durable per-user writable RootFS across tasks, online filesystem growth, and supported file operations while compute remains paused. Sandbox0 Self Host is not available; planned BYOC is not a current substitute for OmniRun’s self-operated option.

Capabilities

OmniRun and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeOmniRunNot 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 filesOmniRunThe reviewed active-sandbox file API and template boot do not establish retained user storage after automatic termination.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSOmniRunNo current durable returning-user RootFS contract is established; user snapshots and hibernation are listed as roadmap.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 stateOmniRunThe current release roadmap places disk hibernation in v0.2 rather than the v0.1 single-node release.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateOmniRunRestoring a prepared template does not prove user-created execution-memory snapshots.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 current roadmap and template-based start do not establish automatic expansion of a retained user RootFS. Compare deployed user-state persistence and growth, independently of base-template start latency.

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 reusable base template is not a saved user machine

OmniRun describes booting prepared template snapshots, while its published v0.2 roadmap places user snapshots, fork, and disk hibernation in future work. A fast clean launch does not prove that packages installed by an earlier agent session return on the next launch. Its self-hosted Claude Managed Agents integration also describes clean execution per tool call; evaluate that integration separately from a continuously active sandbox.

Source 1 ↗Source 2 ↗

Idle deletion needs an explicit export policy

The sandbox guide describes automatic termination after 15 inactive minutes, with an exemption for vaultInject-enabled instances. Applications should export durable outputs and verify effective timeout settings before depending on a quiet environment remaining available. A running shell and file API do not themselves establish a saved RootFS after termination.

Source 1 ↗

Managed pricing and self-operated costs have different boundaries

The managed service has no public numeric rate card or self-service tier: limits and price are negotiated. Its dashboard illustration is an example quota, not an advertised allowance. A zero license fee for the self-operated software still leaves host capacity, storage, upgrades, and operations to the team.

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.

OmniRun

Pricing reviewed 2026-10-18

Managed sandboxes
Price and limits agreed per client or team
No public numeric rate card or self-service plan.
Rate source ↗
Self-operated software license
Apache-2.0; infrastructure billed separately
Verify source availability and release. Host, storage, networking, and operating costs remain the operator’s responsibility.
Rate source ↗

Obtain terms covering resource shape, session timeouts, concurrency, region, retained data, egress, and support. No managed savings figure can be derived from an unpublished quote. Sandbox0’s hosted offer does not imply the same self-operation option.

Check current OmniRun 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 OmniRun

  1. Record template, timeout and idle behavior, and any desktop, vault, network, or managed-agent integration; export state before termination.
  2. Rebuild needed tools in a Sandbox0 template and port the application’s command, file, credential, and service interfaces.
  3. Validate return visits, system changes, growing outputs, paused file operations, and total hosted costs against the actual managed quote or self-operated infrastructure budget.

Common questions

Is Sandbox0 a drop-in OmniRun 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 ↗omnirun.io/llms-full.txt ↗omnirun.io/open-source ↗omnirun.io/docs/sandboxes ↗omnirun.io/managed ↗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 →