ASAgentStead Docs

Choose Your Runtime

Compare the four runtimes available on agentstead.dev and pick the right one for your workflow.

Not sure which to pick? Start with OpenClaw — it's the broadest option and works well for almost every workflow.

Runtimes at a glance

Runtime details

Best for: General-purpose use, coding with Claude Code / Codex / OpenCode, browser-heavy workflows.

  • Full remote desktop with browser tooling pre-installed
  • Largest default workspace allocation
  • Runtime token: Gateway Token
  • Recommended default for new users

Best for: Rust-based agent workflows with an API-oriented runtime surface.

  • Desktop environment included
  • Provider key not required at deploy time
  • Runtime token: API Key
  • Good choice if you prefer Rust tooling over Node-based runtimes

Best for: Minimal footprint — Zig-based runtime with desktop packaging.

  • Lean runtime, lower overhead
  • Desktop included around the core runtime
  • Runtime token: Session Token
  • Choose this if you specifically want the Zig execution model

Best for: Rust performance-oriented workloads.

  • Desktop included
  • Runtime token: Session Token
  • May be rollout-gated depending on your account — check the deploy page for availability

What every runtime includes

Every runtime on agentstead.dev ships with the same platform layer:

Public URL + TLS

A dedicated HTTPS endpoint, always on.

Credentials panel

Username, password, and runtime token in one place.

Persistent workspace

Files survive restarts — no re-cloning on every boot.

Logs & health

Streaming logs and restart controls from the dashboard.

Workspace backups

On-demand and scheduled full workspace backups.

Tailscale networking

Optional private WireGuard access to your agent.

Resource profile

All runtimes currently deploy with the same hardware profile:

ResourceAllocation
RAM4 GB
vCPU2
WorkspacePersistent (size varies by runtime)
Network1 Gbps uplink

Switching runtimes

There's no in-place conversion between runtime types. To switch:

  1. Deploy a new agent of the target type
  2. Migrate any data you need from the old workspace
  3. Destroy the old agent when you're done

Go deeper

Each runtime has a dedicated guide covering startup behaviour, configuration, API key injection, and tool setup: