# Zo Computer

Official pricing: https://www.zo.computer/pricing  
Category: dev-env · Isolation: container

## Pricing regimes (raw)

- **Zo Basic subscription, one always-on computer (4 cores / 32 GB) - $18/mo** (sizes): basic 4 vCPU/32 GiB $0.024658/h [alt]
- **Zo Pro subscription, one always-on computer (16 cores / 128 GB) - $64/mo** (sizes): pro 16 vCPU/128 GiB $0.087671/h [alt]
- **Zo Ultra subscription, one always-on computer (64 cores / 512 GB) - $200/mo** (sizes): ultra 64 vCPU/512 GiB $0.273973/h [alt]
- **N Basic workspaces sized to peak concurrency (part-time use)** (pool) [alt]

## Features

Yes: snapshots, persistent disk, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, root, browser, browser automation API, preinstalled agents, SSH, HTTPS preview URLs, custom domains, raw TCP inbound, Python, Node.js, scheduled runs (cron), MCP server, hosted agent API (their own agent), inbound access rules

No: memory snapshots, fork/clone, pause/resume, volumes, idle auto-stop, your own Docker/OCI image, full VM (own kernel), nested virtualization, desktop GUI, computer-use API, code interpreter, browser + desktop control, anti-bot stealth, CAPTCHA solving, public IPv4, egress allowlist, static egress IP, private networking, self-hosting, BYOC, open source, SOC 2, HIPAA, SSO, GPU, arm64, Windows, macOS, GPU desktop, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, EU data residency, automatic snapshots, snapshots on demand, agent harness API, gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy, secret proxy for any API, extra volumes, shared volumes

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/zo-computer.json → feature_evidence

## Caveats

- Personal cloud computer + AI agent, not a sandbox API: one always-on server per workspace; all modes flagged 'alt'. Fleet = N separate workspace subscriptions (not a documented use).
- Flat monthly subscription: no per-hour billing, so part-time use pays the full $18/$64/$200.
- Plan compute is large for the price (Basic 4 cores/32 GB) but whether cores/RAM are dedicated or caps on shared Modal/Daytona hardware is not stated.
- Included AI credits ($10/$40/$100) cover AI usage only, not compute; not modeled as included_usd.
- Isolation not stated; compute providers are Modal and Daytona (security page) -> modeled as container.
- Zo restarts periodically for updates/snapshots even on paid plans; only registered services come back automatically.
- Storage beyond 100 GB ('100GB+') unpriced; egress, IPv4, regions, compliance undocumented.
- Free Trial: 14 days, limited resources, sleeps when idle, then recovery-only (not usable as a free tier).
- Teams pricing is now published (was 'no prices' in the regime file): Ultra $200/mo for the shared company computer + $18/person/month (each teammate's own Basic) = per-seat cost. 'No contracts, no per-user platform fees'.
- Multiple workspace subscriptions on one billing account are possible (docs mention it), which weakly supports the fleet-basic mode, but paid media is then blocked.

## How this provider charges


Zo is a **personal cloud computer**. Each workspace gets one always-on Linux server with root, a built-in AI agent (50+ tools), hosting
(zo.space, Sites, Services, custom domains), automations, messaging channels (SMS, email, Slack, Discord, Telegram), an MCP server and an API. It is a
consumer/prosumer product, not a sandbox API. Compute runs on **Modal and Daytona** (security page), and the isolation technology is not stated.

**Launched:** public launch 2025-11-19. Sources: HN "Zo: Intelligent Personal Servers for Everyone", https://news.ycombinator.com/item?id=45979424;
Product Hunt launch late November 2025 (#15 of the day); a co-founder post on 2025-12-04 says "we launched" two weeks earlier. Invite/beta coverage dates
from 2025-06-28 (https://news.ycombinator.com/item?id=44407830).

All figures are **list** prices from https://www.zo.computer/pricing and https://www.zo.computer/guide/billing.md.

## Regime table

| Regime | When it applies | How billed | Numbers | Source URL |
|---|---|---|---|---|
| Free Trial | New workspace | $0. 14-day AI + computer trial. Computer sleeps when idle and has "limited" CPU/memory. After 14 days the computer is not started or renewed (up to 7 one-hour recovery sessions, or a support ZIP) | $0; 100 GB disk; 1 hosted service; no custom domains; limited Zo-funded AI | https://www.zo.computer/guide/billing.md |
| Free + Credits | Free workspace with a positive Credits balance | Metered AI/media, and the $0 plan stays. Credits do **not** extend computer access | Media per request (e.g. Nano Banana 2 $0.42/image, Veo 3.1 Fast $0.48–$1.44, transcription $0.0045/audio-min) | https://www.zo.computer/pricing |
| Free daily media allowance | Free, no Credits balance | Free | 3 eligible image, 1 video, 1 transcription requests per workspace per day | pricing page |
| Basic | Paid, monthly | Flat subscription for one always-on computer | **$18/mo**: 4 cores / 32 GB RAM, 100GB+ disk, $10/mo AI credits, 5 services, 3 custom domains | https://www.zo.computer/pricing |
| Pro | Paid, monthly | Flat | **$64/mo**: 16 cores / 128 GB, $40/mo AI, 10 services, 5 domains, priority support | pricing page |
| Ultra | Paid, monthly | Flat | **$200/mo**: 64 cores / 512 GB, $100/mo AI, 50 services, 10 domains, priority support | pricing page |
| AI beyond included credits | Any plan | Credits (token rates per model; media fixed per request) or BYOK / Claude Code / Codex subscriptions | Per /models | https://www.zo.computer/guide/billing.md |
| Plan changes | Upgrade, downgrade or cancel | Immediate, prorated | n/a | billing guide |
| Storage | All plans | Included | 100 GB ("100GB+" on paid); no add-on price published | pricing page |
| Snapshots | Automatic | Regular filesystem snapshots for recovery, not billed | $0 | https://www.zo.computer/reference/information/security.md |
| Egress / IPv4 / regions | n/a | Not documented | null | none |

No per-hour, usage-based or annual pricing is published, and no enterprise/sales tier either (the Teams page has no prices).

## Gotchas

1. **This is not a sandbox API.** You get one computer per workspace. Parallel isolated environments mean separate workspace subscriptions,
   and fleet use is not documented (ToS/account limits unverified).
2. **Flat monthly fees.** A computer used 8 h/day costs the same as one used 24/7. The upside: $18/mo buys 4 cores / 32 GB always-on, which
   is far below per-hour sandbox rates at high utilisation (if those resources are really dedicated, which is not stated).
3. **Included AI credits are AI-only.** They do not offset compute.
4. **The Free Trial is not a free tier.** After 14 days the computer stops for good (recovery-only) unless you pay, and Credits do not extend it.
5. **Periodic restarts** happen for updates and snapshots, even on paid plans. Only registered Services come back automatically, so ad-hoc processes die.
6. **TCP services are public with no auth in front.** Custom domains must be subdomains (no apex).
7. **No GPU, no custom image, no pause/fork, no SDK/CLI for creating machines.**

## Worked example (4 vCPU / 8 GiB, 50 concurrent × 8 h/day × 22 days = 8,800 h, 30% CPU, 50 GiB snapshots, 100 GiB egress)

**Closest analogue:** 50 always-on Basic workspaces, one per concurrent agent. Each is 4 cores / 32 GB, which covers 4 vCPU / 8 GiB. Hours and CPU
utilisation do not change the bill.

| Line | Calculation | $ |
|---|---|---|
| 50 × Basic | 50 × $18 | 900.00 |
| 8,800 h, 30% CPU | Flat fee; always-on anyway | 0 |
| 50 GiB state | Within the included 100 GB disk per workspace; automatic snapshots free | 0 |
| Egress 100 GiB | Not documented | unknown (null) |
| **Total** | | **$900 / month** (+ AI usage beyond the 50 × $10 = $500 of included AI credits) |

Effective rate: $900 / 8,800 h = $0.102 per 4/8-hour used (or $0.0247/h if counted over 730 h always-on).
Alternatives: 50 × Pro = $3,200 (overkill). A hypothetical 4 × Ultra ($800, 256 cores) running agents as processes on shared machines is
cheaper, but those agents would not be isolated sandboxes, so this analogue is not recommended.
Caveat: whether running 50 workspaces as a fleet is allowed, and whether "4 cores" are dedicated, is unverified.