CI and test infrastructure / Buildkite hosted agents alternative

Buildkite hosted agents alternative

Buildkite combines a pipeline control plane with hosted or customer-operated build agents. Hosted agents start clean and are destroyed after each job. A returning AI agent’s tools and mutable system state need a separate persistent execution lifecycle.

Start buildingCompare capabilities ↓

Reviewed 2026-10-19 · Published by the Sandbox0 team

When Buildkite hosted agents fits

Choose Buildkite for dynamic pipelines, test tooling and flexible hosted/self-managed runner fleets. Its macOS hosted machines and enterprise delivery controls may be core requirements for your builds.

When to consider Sandbox0

Consider Sandbox0 for a persistent compatible Linux worker under a retained orchestration layer. You still need the Buildkite agent integration, job isolation, artifacts, credentials and platform-specific build machines.

Capabilities

Buildkite hosted agents and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeBuildkite hosted agentsNot 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 filesBuildkite hosted agentsSelected caches/artifacts can be retained; hosted execution environments are ephemeral.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSBuildkite hosted agentsHosted agents begin clean and are destroyed after each job.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 stateBuildkite hosted agentsPipeline controls do not establish arbitrary machine checkpoint/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 stateBuildkite hosted agentsNo arbitrary hosted-agent RAM checkpoint is established.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.

Buildkite’s hosted cache/image tools help fresh jobs reuse selected state. They do not establish automatic online growth of a whole durable agent RootFS across hosted jobs. A self-managed agent can choose its own persistent backend.

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.

Retained artifacts and control-plane records are useful outside job execution, but they are not full supported-RootFS file operations on a paused guest.

See paused file operations →

Hosted agents intentionally begin clean

Each hosted job receives an isolated ephemeral environment destroyed after execution. Caching can accelerate a later job, while a custom image supplies its initial tools; neither means all changes to the previous machine’s operating system are retained.

Source 1 ↗

Per-vCPU minutes scale with shape

The Pro hosted price table uses $0.004 per Linux vCPU-minute and $0.02 per Mac vCPU-minute. A two-vCPU Linux Small therefore lists $0.008 per minute; a six-vCPU M4 Medium lists $0.12. Usage is metered to the second with no minimum or rounding.

Source 1 ↗

The control plane and hosted compute are separate

Pro’s platform price is per active user. Its 4,000 included Linux vCPU minutes are weighted compute units, not 4,000 minutes on any shape and not a macOS allowance. Agents, test executions, artifacts and other services can have separate limits or charges.

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.

Buildkite hosted agents

Pricing reviewed 2026-10-19

Pro platform
$30 per active user/month
Current plan includes 4,000 Linux vCPU minutes/month; other plan limits apply.
Rate source ↗
Pro Linux hosted compute
$0.004 per vCPU-minute
Linux AMD64 reference shapes; metered by the second without rounding.
Rate source ↗
Pro Mac hosted compute
$0.02 per vCPU-minute
M4 Medium 6 vCPU example is $0.12 per minute, separate from Linux allowance.
Rate source ↗

Platform subscription and execution usage must both be budgeted. ARM64 and larger Linux shapes have separate enterprise availability. A self-hosted agent’s infrastructure and storage remain your responsibility; per-vCPU prices do not prove equal throughput across hardware.

Check current Buildkite hosted agents 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 Buildkite hosted agents

  1. Inventory hosted/self-managed fleets, platform seats, shapes, vCPU-minute allowance and pipeline/test requirements.
  2. Retain orchestration and integrate compatible persistent Linux work with Sandbox0 using explicit cleanup and credential boundaries.
  3. Validate retained system tools, RootFS growth, paused files and total platform, compute, artifacts and model charges.

Common questions

Is Sandbox0 a drop-in Buildkite hosted agents replacement?

Evaluate API compatibility and product scope. These products serve different core workflows, so a change may require a different architecture.

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 ↗buildkite.com/docs/agent/buildkite-hosted ↗buildkite.com/pricing/ ↗Sandbox0 RootFS capacity →

Related comparisons

Apple Xcode Cloud vs Sandbox0 →Artillery vs Sandbox0 →AWS CodeBuild vs Sandbox0 →Azure Pipelines vs Sandbox0 →Bitbucket Pipelines vs Sandbox0 →Blacksmith (GitHub Actions runners) vs Sandbox0 →