Agent sandboxes / Tencent Cloud CubeSandbox alternative

A Tencent Cloud CubeSandbox alternative

CubeSandbox is a self-operated RustVMM/KVM sandbox stack with real memory/filesystem snapshots, rollback, cloning and optional cross-node recovery through S3. Compare its backend choices and infrastructure ownership with Sandbox0 Cloud’s managed, growing writable RootFS.

Start buildingCompare capabilities ↓

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

When Tencent Cloud CubeSandbox fits

CubeSandbox fits teams that want an Apache-licensed KVM sandbox stack, control of deployment and storage, and snapshot/clone workflows. Its SDK surface is E2B-aligned, but individual extensions and current SDK routing need validation rather than assuming every E2B client can target it unchanged.

When to consider Sandbox0

Consider Sandbox0 Cloud when you want the runtime operated as a hosted service with regional object-backed writable RootFS, growth-enabled online expansion and supported paused-file access. KVM-specific kernel, nested virtualization and hardware requirements must be tested against gVisor before migration.

Capabilities

Tencent Cloud CubeSandbox and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeTencent Cloud CubeSandboxThe idle-timeout policy is configurable, including disabled timeout; omitted timeout remains server-controlled.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 filesTencent Cloud CubeSandboxIndependent S3 volumes survive teardown; VM pause/snapshot packages have separate lifecycle/backend rules.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSTencent Cloud CubeSandboxFull-state snapshots include VM filesystem and memory, beyond only a workspace folder.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 stateTencent Cloud CubeSandboxPause/resume preserves full execution state; the default idle-timeout action is kill unless configured otherwise.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateTencent Cloud CubeSandboxSnapshots and clone preserve memory and filesystem; cross-node restoration has compatibility and backend gates.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.

CubeSandbox documents template-selected writable layers, full-state snapshot packages and independent S3 volumes. Its cross-node S3 backend also needs a configured local WAL/cache image. Those features do not themselves establish free-space-triggered automatic growth of a running VM’s whole RootFS; the reviewed guide explicitly separates snapshot-package migration from a live VM 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.

See paused file operations →

Memory and filesystem snapshots are first-class features

CubeSandbox snapshots capture the full running state, and clone creates independent children. Pause/resume retains memory and filesystem while releasing CPU/RAM. Idle timeout defaults to kill unless pause is selected; the server decides an omitted timeout. Choose the lifecycle policy deliberately before treating a workload as persistent.

Source 1 ↗Source 2 ↗

Cross-node recovery requires a specific backend and a ready package

The default XFS backend keeps pause/snapshot packages on the originating node. Cross-node restore requires an S3 backend chosen when the template is built, an uploaded ready package and compatible destination kernel/CPU. The scheduler prefers the origin when it can run the workload. The guide distinguishes a migrating rootfs/memory/metadata package from moving a live VM disk.

Source 1 ↗

Independent S3 volumes are separate from VM snapshots

S3 volumes can outlive a killed sandbox and be reattached across nodes. They are not rewound by restoring a VM snapshot: the restored machine sees current volume contents. Host mounts have a separate node-local availability boundary. Coordinate data versions if the application expects a single consistent point across VM state and external mounted data.

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.

Tencent Cloud CubeSandbox

Pricing reviewed 2026-10-18

CubeSandbox software license
Apache License 2.0; no published license fee
This is open-source software, not a free hosted VM entitlement.
Rate source ↗
Self-operated infrastructure
Your provider and operating costs
Include KVM hosts, storage, local WAL/cache, network, control plane and operations.
Rate source ↗
Managed Tencent Cloud Agent Runtime
Separate product and billing contract
Do not substitute its cloud SKU rates for the cost of your own CubeSandbox deployment.
Rate source ↗

The reviewed repository does not publish a universal hosted CubeSandbox vCPU-hour price. No numeric cloud invoice is invented. Open-source licensing and E2B-aligned APIs do not remove infrastructure charges or operational work; Sandbox0 Cloud is a separate hosted service and is not a self-hosted offer.

Check current Tencent Cloud CubeSandbox 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 Tencent Cloud CubeSandbox

  1. Inventory template backend, local versus remote-ready snapshots, CPU/kernel compatibility, idle-timeout policy, host mounts and independently versioned volumes.
  2. Export required system/user data and recreate compatible images, tools, credentials and network policy on Sandbox0; keep KVM-only dependencies on a suitable platform if necessary.
  3. Validate repeated lifecycle transitions, whole-RootFS growth, committed-state recovery, paused-file behavior and the full infrastructure/operations versus hosted-service cost.

Common questions

Is Sandbox0 a drop-in Tencent Cloud CubeSandbox 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 ↗cubesandbox.com/guide/snapshot-rollback-clone.md ↗cubesandbox.com/guide/lifecycle.md ↗cubesandbox.com/guide/cross-node-snapshot.md ↗cubesandbox.com/guide/s3-volume.md ↗cubesandbox.com/guide/persistent-storage.md ↗github.com/TencentCloud/CubeSandbox/blob/master/LICENSE ↗github.com/TencentCloud/CubeSandbox ↗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 →