Agent sandboxes / Dedalus Labs alternative

A Dedalus Machines alternative for the filesystem and process lifecycle you need

Dedalus Machines already separates storage from compute and retains the root filesystem, including installed system tools, across sleep/wake. Sleeping stops processes rather than preserving their RAM. Compare growth policy, access while sleeping, and storage accounting instead of assuming it only keeps a workspace.

Start buildingCompare capabilities ↓

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

When Dedalus Labs fits

Dedalus fits agents that need a full guest kernel, root access, systemd, persistent system files, autosleep, and typed machine clients. Its price page advertises larger machine and storage allowances on Pro.

When to consider Sandbox0

Consider Sandbox0 for automatic online growth of the durable complete RootFS and supported file operations while compute stays paused. Check kernel-dependent tools and process-continuity requirements against the actual runtime before migration.

Capabilities

Dedalus Labs and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeDedalus LabsAutosleep normally follows five idle minutes and can be changed or disabled. Explicit sleep cancels active executions and closes SSH; retained files have a separate lifecycle.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 filesDedalus LabsRoot-filesystem data persists across sleep/wake, with runtime/temporary exclusions.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSDedalus LabsThe current reference includes /root and system filesystem paths, not only a workspace.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 stateDedalus LabsSleep/wake retains files and identity; stopped processes do not 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 stateDedalus LabsSleep stops processes; wake does not restore their memory or CPU state.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.

Dedalus advertises Pro storage expansion, but the reviewed lifecycle/pricing sources do not establish automatic online growth of the complete mounted RootFS while the agent keeps running. Compare the actual expansion workflow and provisioned snapshot meter.

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 →

Root persistence does not imply memory continuation

The current reference explicitly retains root-filesystem files across sleep/wake, excluding runtime filesystems and temporary state. Sleep ends running processes and SSH sessions; wake retains identity and files but starts with fresh execution state. This is a real complete-system persistence option, with a different process boundary from memory checkpointing.

Source 1 ↗Source 2 ↗

Autosleep and explicit sleep have different activity rules

Machines normally sleep after five idle minutes, and autosleep can be changed or disabled. Explicit sleep cancels active executions and closes SSH; requests needing the machine wait for sleep and can then wake it. Design long tasks and state exports around those actual activity rules.

Source 1 ↗

Storage expansion and stored capacity have their own costs

The price page includes 10 GiB per Hobby machine with no expansion and 20 GiB per Pro machine with advertised expansion. CPU/RAM run by the second, while storage beyond the allowance is billed from provisioned storage snapshots even during sleep. Advertised storage expansion does not itself specify a live RootFS autogrow contract.

Source 1 ↗

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 $65.00 / month

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

Dedalus Labs

Hobby · usage billing

$108.80

$0.14904/hour before monthly allowances

Sandbox0

Pay as you go · no monthly plan fee

$43.80

$0.06/hour · 59% lower in this example

Uses the hourly CPU and RAM meters with second-level billing. Hobby’s $20 credit is one-time and excluded; Pro has separate monthly credit and plan terms. Storage is priced independently.

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

Calculation and billing assumptions

Dedalus Labs: $108.7992 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 Dedalus Labs rate source ↗Sandbox0 rates ↗

Dedalus Labs

Pricing reviewed 2026-10-18

CPU
$0.04536 per vCPU-hour
Current hourly rate view; billed by the second while active.
Rate source ↗
RAM
$0.01458 per GiB-hour
Rate source ↗
Additional provisioned storage
$0.0001 per GiB-hour
Beyond included allowance; storage remains allocated during sleep.
Rate source ↗
Pro
$20 per month including $20 monthly compute credit
20 GiB per machine included; monthly compute credits do not roll over.
Rate source ↗

Hobby’s $20 grant is one-time, not monthly. Pro compute credit is included rather than an extra $20 charge on top of the same credit spend. Confirm effective expansion behavior and provisioned-storage accounting; vendor comparison-calculator savings are not used as an independent result.

Check current Dedalus Labs 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 Dedalus Labs

  1. Export required root files, installed tools, image/service configuration, and data before deleting the source machine.
  2. Port compatible tools, executions, files, credentials, service access, and sleep/activity policy to Sandbox0.
  3. Validate required system paths, growth during long tasks, paused file access, fresh-process versus memory restore, and active compute plus retained-storage cost.

Common questions

Is Sandbox0 a drop-in Dedalus Labs 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.dedaluslabs.ai/llms.txt ↗docs.dedaluslabs.ai/dcs/dm/lifecycle.md ↗www.dedaluslabs.ai/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 →