Agent sandboxes / Novita AI Agent Sandbox alternative

A Novita sandbox alternative for an expanding agent environment

Novita preserves filesystem and memory state through pause/resume, and its long-running mode lifts the normal one-hour timeout cap. Compare disk capacity, idle policy, and independent volume retention alongside the compute rate.

Start buildingCompare capabilities ↓

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

When Novita AI Agent Sandbox fits

Novita fits agents using its code-interpreter or sandbox SDKs and memory-preserving pause. It also offers long-running mode and independently mounted volumes in supported v2 regions; evaluate the actual region and lifecycle configuration.

When to consider Sandbox0

Consider Sandbox0 when the writable environment needs automatic online RootFS expansion and persistent file access without resuming compute. Packages, configuration, and outputs can remain together across successful filesystem checkpoints.

Capabilities

Novita AI Agent Sandbox and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeNovita AI Agent SandboxLong-running metadata lifts the default one-hour timeout cap; configured idle policy still applies.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 filesNovita AI Agent SandboxFiles remain through pause; deletion and independent volumes have different lifecycles.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSNovita AI Agent SandboxThe same sandbox filesystem and memory state return on resume.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 stateNovita AI Agent SandboxPause releases active compute and retains sandbox state.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateNovita AI Agent SandboxRunning processes and variables persist; network connections can be interrupted.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.

Novita’s public pricing lists a 20 GB running sandbox disk allowance. Independent volumes can hold larger or shared datasets, but the reviewed docs do not establish automatic online expansion of the sandbox’s writable root disk.

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.

Novita documents auto-resume through lifecycle settings, including an SDK file-read example after pause. Verify whether a call wakes its sandbox when modeling idle cost; Sandbox0’s supported paused Files API path explicitly leaves compute paused.

See paused file operations →

Long-running mode must be selected deliberately

Novita normally caps a requested timeout at one hour. Setting metadata long_running to the string true allows longer requested timeouts, with documented multi-day examples. It does not disable a configured idle timeout. Carry both settings into a migration plan.

Source 1 ↗

Pause keeps memory, while deletion needs separate storage

Novita pause preserves files, variables, and running processes, with network connections interrupted. Killing a sandbox removes its filesystem; independent beta volumes outlive it and can mount into later sandboxes in supported regions. A retained snapshot is another separate artifact.

Source 1 ↗Source 2 ↗Source 3 ↗

Compare the disk allowance and retained-storage bill

Novita publishes 20 GB of running sandbox storage and a 60 GB account allowance for persistent storage. Retained data above that allowance is billed while paused. Sandbox0’s retained incremental storage also remains billable after compute stops; online growth concerns capacity, not free data retention.

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.

Same size. Lower compute cost.

2 vCPU · 4 GiB RAM · 730 running hours / month

Save $41.35 / month

The same CPU and memory allocation, with runtime aggregated across sessions. Monthly allowances and credits below are applied to this workload.

Novita AI Agent Sandbox

Agent Sandbox usage billing

$85.15

$0.11664/hour before monthly allowances

Sandbox0

Pay as you go · no monthly plan fee

$43.80

$0.06/hour · 48% lower in this example

Uses allocated sandbox CPU and RAM while running. One-time and promotional vouchers are excluded; running and paused storage allowances are priced separately.

Compute and stated plan fees only. Storage, network, taxes and one-time credits are excluded; hardware and throughput can differ.

Calculation and billing assumptions

Novita AI Agent Sandbox: $85.1472 compute after 0 included vCPU-hours and 0 included GiB-hours, plus $0.00 plan fee, minus $0.00 applied monthly credit.

Sandbox0: 4 GiB × 730 hours × its published compute rate. CPU uses the Cloud allocation shown on the pricing page.

Includes compute and the selected plan. Excludes storage, network, taxes, one-time grants and negotiated discounts. Assumes this workload can use the full listed monthly allowances. CPU and memory counts do not guarantee equal hardware or throughput.

Checked Novita AI Agent Sandbox rate source ↗Sandbox0 rates ↗

Novita AI Agent Sandbox

Pricing reviewed 2026-10-18

CPU
$0.0000098 per vCPU-second
Rate source ↗
Allocated memory
$0.0000032 per GiB-second
Rate source ↗
Paused persistent storage above account allowance
$0.00009 per GB-hour
The account allowance is 60 GB; the published running sandbox allowance is 20 GB.
Rate source ↗

CPU and RAM are billed by the second while running. Promotional credits have eligibility and expiry conditions and do not change the rate card. Confirm any independent-volume pricing and region support separately.

Check current Novita AI Agent 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
Understand storage billing →

Before migrating from Novita AI Agent Sandbox

  1. Record long_running metadata, idle timeout, on-timeout action, auto-resume, and the region of each volume.
  2. Export retained sandbox data and independent volumes before any kill operation; recreate secret and ingress policies.
  3. Validate the agent across a long task, a filesystem checkpoint, a paused file edit, and any required explicit memory restoration.

Common questions

Is Sandbox0 a drop-in Novita AI Agent 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 ↗docs.novita.ai/guides/sandbox-long-running ↗docs.novita.ai/guides/sandbox-pause-resume ↗docs.novita.ai/guides/sandbox-volume ↗docs.novita.ai/guides/sandbox-snapshot-create ↗docs.novita.ai/guides/sandbox-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 →