Agent sandboxes / Railway alternative

Compare Railway sandboxes and a persistent agent computer

Railway now has a sandbox VM product with disk checkpoints and forks, alongside its application deployment platform. Use the sandbox-specific price and lifecycle when evaluating agent execution; its ordinary service rates describe another workload.

Start buildingCompare capabilities ↓

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

When Railway fits

Railway fits teams that want agent sandboxes beside existing Railway services, private networking, templates, and project controls. Its checkpoint workflow can prepare environments for repeated code-execution tasks.

When to consider Sandbox0

Consider Sandbox0 when a per-user agent should keep one writable RootFS across tasks and expand it automatically, with compute pauses and persistent file operations independent of a running VM.

Capabilities

Railway and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeRailwayHobby and Pro can disable idle destruction; active command and keepalive behavior still matters.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 filesRailwayNamed disk checkpoints outlive the source sandbox and boot later sandboxes.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSRailwayCheckpoint and fork preserve the sandbox filesystem, with a fresh boot.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 stateRailwayThe reviewed workflow creates or forks from saved disk; idle expiry destroys the source VM.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateRailwayCheckpoint and fork restore files, not running processes or 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.

Railway checkpoints retain a disk and can seed another VM. The reviewed sandbox guide does not establish automatic online expansion of its writable root filesystem; ordinary service volume limits should not be substituted for sandbox root-disk capacity.

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 disk checkpoint can outlive the source sandbox

Railway checkpoints save a running sandbox disk and become bootable reusable state. Forks also copy the disk into a new sandbox. Both boot fresh processes, so a retained filesystem is different from preserving an interpreter or open process state.

Source 1 ↗

Idle teardown needs an explicit retention plan

Railway destroys a sandbox after its idle timeout unless the paid workspace disables it. Active exec or SSH work can defer teardown, but an unrelated background process does not necessarily keep the sandbox alive. Save the required checkpoint before destruction and restore the desired configuration on each new fork.

Source 1 ↗

Sandbox memory is charged at the VM rate

Railway’s sandbox VM rate is $50 per GB-month of memory in use and $50 per vCPU-month of CPU actually consumed, using a 30-day month. Memory includes operating-system and filesystem-cache use. Those differ from Railway service rates and from Sandbox0’s configured-memory compute meter.

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.

Railway

Pricing reviewed 2026-10-18

Sandbox CPU actually consumed
$0.00001929 per vCPU-second
The published monthly equivalent is $50 per continuously used vCPU for a 30-day month.
Rate source ↗
Sandbox memory in use
$0.00001929 per GB-second
Includes OS and filesystem cache; published equivalent is $50/GB-month while running.
Rate source ↗
Sandbox egress
$0.05 per GB
Rate source ↗
Hobby minimum spend
$5/month including $5 of usage
Rate source ↗
Pro minimum spend
$20/month including $20 of usage
Rate source ↗

Use the Sandboxes & VMs section of the Railway rate card. Ordinary deployment CPU, memory, volumes, and object-storage rows are separate. Usage above included plan credit is additional; idle memory and background CPU can remain billable.

Check current Railway 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 Railway

  1. Record checkpoint names, source regions, templates, file changes, and idle settings before destroying a Railway sandbox.
  2. Port private-network dependencies, domains, environment variables, SSH access, and startup commands; forks do not inherit every source option.
  3. Test growing files, system-package persistence, paused file access, and fresh-process startup on Sandbox0, then compare measured CPU use with the intended memory tier.

Common questions

Is Sandbox0 a drop-in Railway 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.railway.com/sandboxes ↗railway.com/pricing ↗/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 →