agents-sandbox vs Minimal
Choose agents-sandbox for developers who want each coding agent confined to a per-project microVM on their own Linux or Apple Silicon machine; choose Minimal for reproducible, declaratively scoped local sessions for Claude Code, Codex, Gemini and Pi with a supply-chain-audited toolchain. agents-sandbox uses microsandbox microVM isolation and is open source; Minimal uses namespaces or libkrun microVM isolation and is open source.
Key differences between agents-sandbox and Minimal
- Isolation: agents-sandbox uses microsandbox microVM; Minimal uses namespaces or libkrun microVM.
- Licence: agents-sandbox is GPL-3.0; Minimal is Apache-2.0.
- SDKs and interfaces: Only Minimal lists MCP server.
What they have in common
- Both are tools that run on your own machine.
- Both are open source.
- Both list CLI.
agents-sandbox vs Minimal at a glance
| Attribute | agents-sandbox | Minimal |
|---|---|---|
| What it is | GPL-3.0 launcher that gives each project and coding agent its own microsandbox microVM on Linux/KVM or Apple Silicon macOS, with secrets kept on the host. | Open-source CLI that runs coding agents in declarative, isolated sessions built from a minimal.toml and a from-source package catalogue, on Linux namespaces or a libkrun microVM. |
| Type | Local / OS-level | Local / OS-level |
| Isolation | Dedicated microsandbox microVM per project and agent with a "separate microVM and Linux kernel"; the README does not name the hypervisor | Linux: unprivileged user and mount namespaces by default (local-minimald), or a libkrun microVM with --provider local-minvmd; macOS: always a lightweight Linux microVM booted by libkrun on Hypervisor.framework |
| Open source | Yes (GPL-3.0) | Yes (Apache-2.0) |
| Self-host / BYOC | No / No | No / No |
| SDKs | CLI | CLI, MCP server |
| Cold start (vendor claim) | Not documented | Not documented |
| Max session | Not documented | Not documented |
| Persistence | VM is built or reused per project; agent state persists in a /home/dev volume; the project directory is mounted read-write at /workspace, or an isolated VM-internal worktree for some agents | Sessions can be attached, detached and kept running; the session workspace starts as a copy of the project files taken at activation; a /state directory can be shared across tasks via state_key |
| Snapshots / pause | Not documented / Not documented | Not documented / Not documented |
| GPU | Not documented | Not documented |
| Networking | Egress profiles: public (default), private and host extend it, none is allowlist-only "deny-by-default egress, not a complete air gap"; links can open in the host browser for allowlisted schemes | Package builds get network access only when a dependency requires it and tasks "when necessary"; session network policy is not documented |
| Pricing | Free (GPL-3.0) | CLI is in public beta and free to download for Linux and macOS; "Minimal for Teams" (managed hermetic builds, org-wide policy enforcement, attestation) is marked "Coming Soon" |
| Free tier | Free | Free to download (public beta) |
| Compliance | Not documented | Not documented |
| Regions | Not documented | Not documented |
| Company headquarters | Hamburg, Germany | Not documented |
| Good fit for | Developers who want each coding agent confined to a per-project microVM on their own Linux or Apple Silicon machine | Reproducible, declaratively scoped local sessions for Claude Code, Codex, Gemini and Pi with a supply-chain-audited toolchain |
Verified November 14, 2026 from each provider’s documentation. Cold-start figures are vendor claims, not our measurements.
Which should you choose?
Reasons to choose agents-sandbox
- You want one microVM per project rather than one shared sandbox
- You need secrets kept off the guest while the agent still works
- You want configurable egress without a full air gap
Reasons to choose Minimal
- You want the agent environment declared in a file next to the code and rebuilt reproducibly
- You care about supply-chain provenance of the tools inside the sandbox
- You are on Apple Silicon and want a microVM rather than a container
Trade-offs to check first
- agents-sandbox: Five of the seven agent profiles are marked experimental
- agents-sandbox: Hypervisor and boot behaviour are inherited from microsandbox and not documented here
- Minimal: Public beta; Teams and remote execution are marked coming soon; the full source workspace does not yet build on macOS
- Minimal: Linux default sandbox uses user and mount namespaces; the libkrun microVM provider is opt-in there
Both are worth piloting with your own image and region. See the full agents-sandbox profile and Minimal profile, or the complete comparison.
More comparisons
Frequently asked questions
Is agents-sandbox or Minimal better for AI agents?
What is the difference between agents-sandbox and Minimal?
Which is open source, agents-sandbox or Minimal?
How do agents-sandbox and Minimal isolate code?
How does agents-sandbox pricing compare with Minimal?
Can you self-host agents-sandbox or Minimal?
Sources
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). agents-sandbox vs Minimal. agentsandboxes.org. Last reviewed November 14, 2026. https://agentsandboxes.org/compare/agents-sandbox-vs-minimal/