ComfyDeploy’s current pricing page says it no longer accepts new customers. Existing customers’ service and support remain unchanged. Read the source ↗

GPU clouds / ComfyDeploy (existing hosted customers) alternative

ComfyDeploy alternative for portable persistent agent work

ComfyDeploy provides ComfyUI workflow deployment, team interfaces and GPU-backed execution. Its current pricing page says it no longer takes new customers while existing service and support remain unchanged. The 2025 open-source announcement and ongoing hosted-customer service must be distinguished from new hosted onboarding.

Start buildingCompare capabilities ↓

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

When ComfyDeploy (existing hosted customers) fits

Existing customers can evaluate continued ComfyUI execution and support under their actual terms. The open-source platform is another deployment path with its own GPU, storage and operating costs; it is not the same as a new hosted account.

When to consider Sandbox0

Consider Sandbox0 for compatible Linux agents that orchestrate generation workflows and retain their own installed tools and growing data across tasks. Keep GPU rendering and required ComfyUI execution on suitable infrastructure.

Capabilities

ComfyDeploy (existing hosted customers) and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeComfyDeploy (existing hosted customers)Not 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 filesComfyDeploy (existing hosted customers)Private model storage and reusable workflow environments retain selected artifacts; current hosted access is limited to existing customers.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSComfyDeploy (existing hosted customers)Workflow/model storage does not establish complete mutable agent-OS retention.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 stateComfyDeploy (existing hosted customers)The cited workflow platform does not establish a general process-preserving returning-agent pause/resume contract.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateComfyDeploy (existing hosted customers)Reusable workflow/model artifacts do not establish arbitrary process/VRAM checkpoint recovery.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.

ComfyDeploy’s reusable machine/workflow configuration and private model storage address generation workloads. They do not establish automatic online low-space growth of a general agent’s entire durable RootFS.

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.

Workflow files, model storage and generation outputs have their own interfaces. The cited service does not establish arbitrary reads/mutations throughout a paused full OS filesystem with guest compute remaining stopped.

See paused file operations →

New-customer availability changed

The current hosted pricing notice says new customers are no longer accepted and existing service/support remains unchanged. The September 2025 announcement described reopening the platform source and supporting hosted users. Use the current notice for onboarding rather than assuming the older future-customer PAYG invitation still applies.

Source 1 ↗Source 2 ↗

Workflow environments and model storage have a specific scope

The docs describe machine configurations, private model storage, cloud ComfyUI workspaces and API deployments. These preserve reusable workflow/model artifacts but do not establish arbitrary whole-OS retention or process/VRAM checkpoints for a general-purpose agent.

Source 1 ↗Source 2 ↗

Use the existing account’s effective plan

The public page still lists GPU usage references and a Business starting amount. Confirm the effective existing contract, credit treatment, selected GPU and retained resources before adding fees. Open-source deployment requires its own infrastructure quote rather than inheriting hosted plan pricing.

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.

ComfyDeploy (existing hosted customers)

Pricing reviewed 2026-10-19

New hosted account
Not accepting new customers
Existing service/support remains unchanged under the current notice.
Rate source ↗
Published H100 usage reference
$8.4168 per GPU-hour
Existing-customer table reference; verify effective plan/credits and configuration.
Rate source ↗
Published T4 usage reference
$0.6480 per GPU-hour
Different hardware/profile; not equivalent inference/rendering throughput.
Rate source ↗
Business starting reference
From $998/month
Confirm included credits and additional usage rather than adding every fee unconditionally.
Rate source ↗

The live pricing page supersedes older onboarding descriptions. Published values are references for evaluating an existing account, not an available new signup offer. Confirm GPU/time billing, credit allocation, model storage and other resources; independent open-source hosting has separate costs.

Check current ComfyDeploy (existing hosted customers) 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 ComfyDeploy (existing hosted customers)

  1. Confirm existing-account status, effective plan/credits, workflow definitions, private models/storage and the supported execution path.
  2. Export needed portable files and keep GPU rendering/workflow execution on suitable infrastructure while moving compatible Linux orchestration.
  3. Test result handoff, persistent tools, growth, paused files and full existing-hosted or independent GPU/storage plus Linux runtime costs.

Common questions

Is Sandbox0 a drop-in ComfyDeploy (existing hosted customers) 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 ↗app.comfydeploy.com/pricing ↗www.comfydeploy.com/blog/re-open-sourcing-comfydeploy ↗docs.comfydeploy.com/docs/introduction ↗www.comfydeploy.com/ ↗Sandbox0 RootFS capacity →

Related comparisons

Azure GPU VMs (NC_A100_v4 reference) vs Sandbox0 →Cerebrium Serverless Compute vs Sandbox0 →Crusoe vs Sandbox0 →DigitalOcean GPU Droplets vs Sandbox0 →Fluidstack vs Sandbox0 →Google Compute Engine GPU VMs vs Sandbox0 →