Agent Sandboxesagentsandboxes.org
Sandbox profile · Cloud or model provider service

DigitalOcean Harness Runtime

Answer in brief

DigitalOcean's managed microVM sandboxes that run unmodified agents such as Claude Code, Codex or LangGraph, composed from a YAML environment of adapter, size, permissions, secrets, tools and egress rules. A composable, agent-first microVM service with checkpoint and fork support and consumption-based CPU billing, currently in public preview and listed in one datacenter (RIC1).

By the Agent Sandboxes editors · Updated

Overview

Harness Runtime is the sandbox half of DigitalOcean Managed Agents, which entered public preview for all users on 21 September 2026 alongside Action Gateway, a managed MCP endpoint in front of more than 16,000 tools. Each session gets its own Firecracker microVM with its own kernel. The agent runs inside it unmodified through an adapter: Codex, Claude Code and OpenCode for coding, and LangGraph, CrewAI, Hermes and OpenHands as general frameworks. Harness Runtime translates each agent's native output into one event stream.

The product is composable. An environment is an immutable YAML spec, or a Control Panel form, that combines the adapter (harness), a size from 1 vCPU / 1 GB to 16 vCPU / 32 GB, Agent Skills, secrets, permission rules that allow, ask about or deny shell commands and file paths, MCP and Action Gateway tools, and an egress allowlist. Teams can register their own OCI image as a template, rebased onto one of DigitalOcean's base images. Since 26 October 2026, harness: none boots a bare Ubuntu 24.04 sandbox that you drive with exec, file upload and download, and port forwarding through doctl. Checkpoint, fork and rollback work as they do for other sessions.

Sessions pause after 15 idle minutes by default, keeping memory, files and running processes, and resume on the next input. Checkpoints capture the whole machine, so you can fork a session or roll it back. Billing is per second of active use, with CPU charged on consumption rather than allocation, and pausing stops compute charges. Clients are doctl, the REST API, and PyDo, DoTs and Godo client libraries for Python, TypeScript and Go. A Cuttlefish iOS app lets you approve tool calls from a phone. The availability page lists a single datacenter, RIC1.

Notable

  • One Firecracker microVM per session, with checkpoints, forks and rollback of the whole machine
  • Runs Claude Code, Codex, OpenCode, LangGraph, CrewAI, Hermes and OpenHands unmodified through adapters, or a bare sandbox with harness: none
  • CPU billed on consumption at $0.044 per vCPU-hour, so an agent waiting on a model costs little; up to 100 concurrent sessions per team

DigitalOcean Harness Runtime specifications

AttributeDetails
IsolationOne Firecracker microVM per session with its own kernel; DigitalOcean says isolation is enforced by the hypervisor rather than container namespaces
DeploymentCloud or model provider service
Open sourceNo
SDKs and interfacesPython, JavaScript/TypeScript, Go, CLI, REST API
Cold start (vendor claim)Creation to a first agent response in "under a couple of seconds"; paused sessions resume in about 300 ms according to the product page intro (its FAQ says about 200 ms) (vendor claims)
Max session lengthNo fixed session length documented; sandboxes pause after an idle timeout (default 15 minutes, configurable, or disabled with keep_warm: true) and resume on the next input, exec or port-forward; each session has a monthly active-compute cap of 744 hours
PersistencePausing keeps memory, files and running processes with compute stopped; checkpoints capture the whole machine for fork and rollback; Persistent Workspaces (public preview, 1 to 100 GiB) save /workspace when a session is removed or checkpointed so a later session can attach it
GPUNot documented
NetworkingOutbound access is unrestricted by default; an egress hostname allowlist confines the sandbox to the listed hosts. Ports reach your machine only through doctl harness-runtime port-forward, and nothing is published to the internet. Bandwidth is 32 to 512 Mbps by size; private VPC connectivity and VPC peering are not supported
PricingBilled per second of active use: CPU $0.044 per vCPU-hour on CPU actually consumed, memory $0.0095 per GB-hour allocated, so the default 2 vCPU / 4 GB sandbox is at most $0.126 per hour; session storage, checkpoints and custom templates $0.05 per GiB-month; public egress $0.01 per GiB; paused sessions incur no compute charges. Optional Pro ($50/month) and Team ($200/month) plans pre-buy usage at a 15% or 20% discount
Free tier$5 promotional credit for new customers; Persistent Workspaces are free during public preview
ComplianceNot documented
RegionsRIC1 only, per the availability page
Company headquartersBroomfield, Colorado, United States · More from the United States

Facts from DigitalOcean Harness Runtime’s public documentation, verified November 14, 2026. Performance numbers are vendor claims.

Strengths and trade-offs

Strengths

  • You want to run an off-the-shelf agent such as Claude Code or Codex in a managed microVM without wrapping it yourself
  • You need allow, ask and deny permission policies, approvals and an egress allowlist declared per environment
  • You want pause, checkpoint, fork and rollback for long-running agent sessions
  • You already use DigitalOcean for databases, storage or inference

Trade-offs

  • The whole product is in public preview, and Persistent Workspaces and Signals are preview features within it
  • Only one datacenter (RIC1) is listed, with no VPC peering or private connectivity
  • Compliance certifications and GPU support are not documented for Harness Runtime
  • Egress is unrestricted unless you set an allowlist

Good fit for: Teams that want to run unmodified coding agents such as Claude Code or Codex in managed microVMs with permission policies, checkpoints and forks, especially if they already use DigitalOcean.

Use cases DigitalOcean Harness Runtime fits

  • Coding agents: Runs Claude Code, Codex or OpenCode unmodified in a Firecracker microVM per session, with allow, ask and deny permission rules, checkpoints and forks.

How DigitalOcean Harness Runtime compares

SandboxCompany HQIsolationOpen sourceSelf-host / BYOCCold start (vendor)Max sessionSnapshots / pauseGPUSDKs
DigitalOcean Harness RuntimeUnited StatesFirecracker microVMNoNo / Not documented<2 s (claim)No fixed limit; idle pause after 15 minYes / YesNot documentedPython, JavaScript/TypeScript, Go, CLI, REST API
AWS Bedrock AgentCoreUnited StatesMicroVM (Runtime)NoNo / Not documentedNot documentedUp to 8 hNot documented / Not documentedNot documentedPython (AgentCore SDK), AWS SDKs, AWS CLI
E2BUnited StatesFirecracker microVMYesYes / YesNot documented1 h / 24 h (Pro)Yes / YesNoPython, JavaScript/TypeScript
DaytonaUnited StatesContainers or VMsNoNot documented / Yes<90 msUnlimitedYes / YesYesPython, TypeScript, Ruby, Go, Java
Fly.io SpritesUnited StatesMicroVMNoNot documented / Not documented0.1–2 s wakeUnlimitedYes / YesNoJavaScript/TypeScript, Go, Python, Elixir
Morph CloudNot documentedVMNoNot documented / Not documented<250 ms (restore)ConfigurableYes / YesNot documentedPython, TypeScript
BlaxelUnited StatesMicroVMNoNot documented / Yes<25 ms resumeUnlimited (standby)Yes / YesNot documentedTypeScript, Python, Go

Head-to-head: AWS Bedrock AgentCore vs DigitalOcean Harness Runtime

See all DigitalOcean Harness Runtime alternatives →

Frequently asked questions

What is DigitalOcean Harness Runtime?
DigitalOcean's managed microVM sandboxes that run unmodified agents such as Claude Code, Codex or LangGraph, composed from a YAML environment of adapter, size, permissions, secrets, tools and egress rules. A composable, agent-first microVM service with checkpoint and fork support and consumption-based CPU billing, currently in public preview and listed in one datacenter (RIC1).
How does DigitalOcean Harness Runtime isolate code?
DigitalOcean Harness Runtime documents its isolation as: One Firecracker microVM per session with its own kernel; DigitalOcean says isolation is enforced by the hypervisor rather than container namespaces.
Is DigitalOcean Harness Runtime open source?
No, DigitalOcean Harness Runtime is not open source.
How much does DigitalOcean Harness Runtime cost?
Billed per second of active use: CPU $0.044 per vCPU-hour on CPU actually consumed, memory $0.0095 per GB-hour allocated, so the default 2 vCPU / 4 GB sandbox is at most $0.126 per hour; session storage, checkpoints and custom templates $0.05 per GiB-month; public egress $0.01 per GiB; paused sessions incur no compute charges. Optional Pro ($50/month) and Team ($200/month) plans pre-buy usage at a 15% or 20% discount Check the official pricing page for current rates.
Does DigitalOcean Harness Runtime have a free tier?
$5 promotional credit for new customers; Persistent Workspaces are free during public preview.
How long can DigitalOcean Harness Runtime sessions run?
No fixed session length documented; sandboxes pause after an idle timeout (default 15 minutes, configurable, or disabled with keep_warm: true) and resume on the next input, exec or port-forward; each session has a monthly active-compute cap of 744 hours.
Does DigitalOcean Harness Runtime support snapshots or pause and resume?
Snapshots: yes. Pause and resume: yes. Pausing keeps memory, files and running processes with compute stopped; checkpoints capture the whole machine for fork and rollback; Persistent Workspaces (public preview, 1 to 100 GiB) save /workspace when a session is removed or checkpointed so a later session can attach it.
Can you self-host DigitalOcean Harness Runtime?
DigitalOcean Harness Runtime is offered as a managed service and does not document a self-hosted option.
Which regions does DigitalOcean Harness Runtime run in?
RIC1 only, per the availability page.
Where is DigitalOcean based?
DigitalOcean, the organisation behind DigitalOcean Harness Runtime, is headquartered in Broomfield, Colorado, United States. That is where the company is based, not where sandboxes run; for that, see the regions listed for DigitalOcean Harness Runtime.
What are the alternatives to DigitalOcean Harness Runtime?

Sources

  1. digitalocean.com/products/managed-agents
  2. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/
  3. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/details/features/
  4. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/concepts/architecture/
  5. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/concepts/environments/
  6. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/details/pricing/
  7. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/details/limits/
  8. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/details/availability/
  9. docs.digitalocean.com/products/managed-agents/agent-harness-runtime/how-to/run-sandbox/
  10. docs.digitalocean.com/release-notes/agent-harness-runtime/
  11. digitalocean.com/blog/microvms-for-agents
  12. sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001582961&type=10-K&dateb=&owner=include&count=10

Check before you rely on it. Products change quickly, and this page reflects public documentation as of November 14, 2026. Confirm anything that matters in the provider’s own documentation.

Cite this page

Agent Sandboxes. (2026). DigitalOcean Harness Runtime. agentsandboxes.org. Last reviewed November 14, 2026. https://agentsandboxes.org/sandboxes/digitalocean-harness-runtime/