Frequently asked questions
Cloud Terminal for scalable work.
Everything you need to know before you spin up your cloud terminal — CLI, scale, AI agents, billing, security, regions, migration.
1 General
Is it self-hosted or fully cloud?
Fully cloud-hosted by us. No install, no setup. You sign in and your terminal is ready — running on Cloudflare's edge + Hetzner Singapore/EU/US compute. Persistent across reconnects, scalable to dozens of parallel sessions.
What is devsclouds?
Cloud Terminal for scalable work — your laptop, in the cloud. Real terminal in your browser (xterm.js, full shell, tmux, SSH-class power), persistent sessions across devices, parallel terminals when you need scale, and any AI agent you want: Claude, Codex, Gemini, OpenClaw, aichat, or none. Singapore / EU / US regions. ~20% less than the alternatives.
How does scale work? Can I run more than one terminal at a time?
Yes. Three concrete things. (1) Parallel sessions: every tier supports multiple simultaneous terminals in the same workspace — LITE 4, PRO 8, MAX 16, ULTRA 32+. Run one CLI per task, side by side. (2) Multi-region: pick Singapore (asia-southeast), Germany (EU-Central), Finland (EU-North), or US Virginia (US-East) at signup so the keystroke echo stays sub-50ms. (3) Snapshot & clone: snapshot your container before a risky migration, clone it for a parallel branch experiment, restore from any point in the last 30 days. Compute per box is fixed-size (Hetzner cpx tier per plan) — we do not autoscale a single box; scale is by sessions, regions, and clones.
Do I need to use AI? Can I just use it as a cloud terminal?
Yes, just a cloud terminal works fine. No AI required. You get a real container with a real shell — ship code, run servers, attach to tmux from your phone, share a session with a teammate. The AI integrations (Claude Code, Codex, Gemini, OpenClaw, BYOK keys, 12 channels) are optional layers on top.
Which AI agents work on devsclouds?
All of them. Claude Code, OpenAI Codex, Gemini CLI, OpenClaw, aichat, llm, fabric — if it runs on Linux, it runs here. Bring your own keys (Anthropic, OpenAI, Google) at zero markup, or use our metered AI Credits pool. Switch agents mid-session, run several side by side.
How long does it take to spin up?
Under 60 seconds in the happy path. Sign in with Google, pick a tier, pick a region, and you are dropped into a real shell. No Docker, no YAML, no SSH keys to lose. We keep a pool of warm images so boot is near-instant.
Do I need to be a developer?
Not strictly — but you should be comfortable with a terminal. The browser shell is a real Linux box with bash, git, bun, node, python, vim, tmux, and friends. If you would feel at home in a laptop terminal, you will feel at home here.
How is this different from running a VPS or Codespaces?
It is closer to a Codespaces-style cloud workstation than a raw VPS — persistent state, browser-native UI, zero ops. Versus Codespaces: AI-agnostic (any agent, BYOK), three regions including Singapore, ~20% cheaper, and 100% upstream OpenClaw parity for users who want it.
2 Billing
How is devsclouds priced?
Four tiers from $12/mo yearly to $119/mo: LITE, PRO, MAX, ULTRA. Yearly saves ~17% over monthly. Roughly 20% under comparable cloud-dev offerings, with Singapore/EU/US regions standard. Our /pricing page has the full side-by-side.
What is BYOK and why does it matter?
BYOK = Bring Your Own Key. You paste your own Anthropic, OpenAI, or Google API key; we inject it into your container at boot, encrypted at rest and never logged. You pay the LLM provider direct at zero markup. Optional — if you do not use AI, you do not need a key.
What is the AI Credits pool?
Optional alternative to BYOK. You add credits via Stripe, and we meter token usage against them so you do not need to manage LLM accounts. Useful for team workspaces where you want a single pool.
Can I cancel anytime? What happens to my data?
Cancel anytime from the dashboard on monthly plans. Your instance is paused immediately; your data is retained for 30 days so you can export or re-subscribe. After 30 days, it is permanently deleted per our retention policy.
3 Security & compliance
Are you SOC 2 / HIPAA compliant?
We rely on Cloudflare (SOC 2 Type II, ISO 27001, ISO 27018, PCI DSS, GDPR) and Hetzner (SOC 2 Type II, ISO 27001, ISO 9001, GDPR, TÜV Süd-certified data centers) for all infrastructure. Your data inherits their certifications. We do not yet hold our OWN SOC 2 audit and we do not currently sign HIPAA BAAs — we are transparent about this. See our <a href="/compliance" class="text-emerald-600 hover:underline">compliance page</a> for the full subprocessor list with each provider's certifications.
Where is my data stored?
You pick the region: Singapore (Hetzner SIN data center), EU (Hetzner Falkenstein or Helsinki), or US (Hetzner Ashburn). Edge traffic terminates at Cloudflare's nearest PoP and is proxied to your chosen region; your container, memory, and backups all live in that region. Hetzner is GDPR-compliant and ISO 27001 certified across all three locations.
How do you isolate my data from other tenants?
Every tenant runs in its own Docker container with dedicated named volumes, dedicated memory, and no route to other tenants' disks — never shared infrastructure. MAX and ULTRA add gVisor syscall-level hardening. Your BYOK keys are injected at boot from encrypted storage and never written to shared filesystem.
What about supply-chain attacks on AI skills?
In January 2026, 341 malicious skill packages exfiltrated keys from ~1,800 self-hosted OpenClaw instances (the “ClawHavoc” incident). Signed-skill verification is on our roadmap — the goal is that only skills signed by trusted author keys would install. Until that ships, the same per-tenant container isolation contains any compromised skill to your own workspace. See our blog post for the full post-mortem.
Do you publish a DPA and subprocessor list?
Yes. Our DPA is linked from the footer and the <a href="/compliance" class="text-emerald-600 hover:underline">compliance page</a> lists every subprocessor — Cloudflare, Hetzner, Google, Anthropic, OpenAI, Stripe, WorkOS — with each one's actual certifications and what data they touch.
How is my traffic encrypted?
TLS 1.2+ end-to-end via Cloudflare Universal SSL (LetsEncrypt + Google Trust Services certificates, auto-rotated). Your browser → Cloudflare edge → Hetzner container, all encrypted in transit. At rest, your container disk uses LUKS encryption on the underlying Hetzner volume.
4 Channels
Which 12 messenger channels does devsclouds support?
WhatsApp, Telegram, Slack, Discord, iMessage, Signal, Microsoft Teams, Matrix, Feishu / Lark, Zalo, Nostr NIP-17, and Nextcloud Talk. Each is a two-way adapter for any agent you wire up. Optional — ignore them entirely if you just want a cloud terminal.
How does the iMessage bridge work?
We bridge through a self-hosted Matrix gateway on your Mac, so messages stay end-to-end encrypted on the Apple side. You sign in once with your Apple ID on a Mac you own; devsclouds never sees your Apple credentials.
Can my agent switch channels mid-conversation?
Yes. Conversation memory, preferences, timezone, and language live in the agent, not in the channel. Your supplier starts in Feishu, you reply from iMessage, and the agent maintains the thread across both.
5 Technical
Where is my data stored? Which regions?
You pick home_region at signup: Singapore (asia-southeast), Germany (EU-Central), Finland (EU-North), or US Virginia (US-East). Your container, memory, and backups all live in that region. Pick the region closest to you for low-latency terminal feel.
How latent is the browser terminal?
Sub-50ms RTT to your nearest region for keystroke echo. We run xterm.js over a WebSocket pinned to your container — no proxy hops, no shared multiplexer. It feels like a local terminal once the session is attached.
Does it work offline / on a flaky connection?
Sessions are tmux-backed and persist server-side, so a dropped connection is just a re-attach. You can close the tab on your phone, open it on your laptop an hour later, and pick up exactly where you left off — with the same scrollback and running processes.
Can I use community skills from the OpenClaw Skills Hub?
Yes — OpenClaw is one of the supported agent runtimes and we run the unmodified upstream image, so every community skill is compatible. Optional signed-skill verification is on the roadmap (post-ClawHavoc). You can opt in per-skill, or ignore OpenClaw entirely and use any other CLI agent.
What happens if my container crashes?
We auto-restart on crash with exponential backoff, snapshot your state on every commit, and replay in-flight tasks from the last known-good checkpoint. Daily off-site backups are encrypted and retained 30 days.
Do you support teams, RBAC, and SSO?
MAX and above get team workspaces in 2026. Enterprise adds RBAC, WorkOS SSO (SAML + OIDC), SCIM provisioning, and immutable audit logs. MVP is single-user.
6 Migration
How do I migrate an existing OpenClaw instance?
Export your OpenClaw state (standard tarball), sign up to devsclouds, pick a region, and upload via the dashboard. Your skills, memory, and chat history transfer 1:1 since we run the unmodified upstream image. Typical migration: under 10 minutes.
How do I move my dotfiles / SSH keys / dev setup?
Standard Linux home-dir — clone your dotfiles repo, drop SSH keys in ~/.ssh, install whatever you want with apt/bun/npm/pip. The container persists across sessions, so a one-time setup is exactly that. Reproducible setups via a tracked init script if you prefer.
Will my community skills keep working after migration?
Yes. We run unmodified upstream OpenClaw, so any community skill installs identically on devsclouds. If you opt into signed-skill verification, unsigned skills will refuse to install — you can flip that switch off if needed.
Do you offer migration assistance on paid tiers?
Yes, free on ULTRA and Enterprise. One of our engineers will walk your migration live, verify state parity, and wire up any custom channels you use. MAX and below get written migration docs and email support.