boat vs Zeabur

Zeabur is an app platform. You rent a server by the month, or connect your own, and deploy apps and databases onto it. boat makes a full VM for each agent by API, bills it per second, saves its disk every minute and forks it while it runs. Each sandbox has a desktop and coding agents installed. boat has limits too: plans start at $20 a month, starts per day are capped, and sandboxes run only in Germany, Finland and France.

Where boat stands out

Machines from code, billed per second

boat creates a sandbox from code when an agent needs one and bills it per second while it runs. It saves the disk every minute and at each stop. A stopped sandbox has no compute charge, and boat keeps its disk for at least 14 days. A Zeabur server is bought for a month and bills until you cancel it, busy or idle.

Fork a running sandbox

On boat you fork a sandbox while it runs, or start many from a saved snapshot, each with the same disk. Zeabur backs up volumes and databases on the Dev plan and up, kept for 7 days. It documents no snapshot or clone of a whole server.

Coding agents and a desktop installed

Every boat sandbox has seven coding-agent harnesses installed, including Claude Code, Codex and OpenCode. You drive them from outside with a prompt, steer and events API. Each sandbox also streams a 60fps desktop with Chromium. The optional, paid credentials proxy keeps real AI keys out of the machine. Zeabur's MCP server and Claude Code plugin work the other way: your assistant uses them to deploy apps.

The companies and products

boat

boat.dev runs sandboxes for AI agents: each one is a full Linux VM with its own kernel.

You start a machine with one command or one API call, then SSH in, run Docker, open a desktop, or let a built-in agent (Claude Code, Codex, OpenCode and others) work in it. The disk stays between sessions, and a stopped sandbox costs nothing.

Zeabur

Zeabur is a platform for deploying apps, databases and templates onto servers you rent through it or bring yourself.

You buy a server from a catalog of AWS, GCP, Hetzner, Linode, DigitalOcean, Aliyun, Tencent Cloud and Volcengine machines, at a fixed monthly price, or you connect one you already own. With ZeaburOS (built on K3s) the server becomes a deployment node. Your services run on it as containers, with logs, domains, email and backups in the dashboard. With plain Ubuntu it is a normal VPS that you run over SSH. Zeabur is built for teams that host web apps. This page compares the two where they overlap: an agent on its own Linux machine.

Which use cases each one fits

For each use case, Battleships prices every sandbox provider on the same workload and sorts them from the cheapest monthly bill up. “5th cheapest of 365” means 4 of the other 364 providers run that workload, with every feature it needs, for less. “With limits” means the provider runs it but misses something, so it is listed after every full match. Open a use case for its whole ranking and what each provider misses.

Use case and its workloadboat: price rankZeabur: price rank
Software factory
100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month
A software factory creates and deletes many machines a day. boat bills each one per second, within its start caps. On Zeabur each machine is a server bought for a month; you reuse a few servers instead of creating new ones.
5th cheapest of 365
$634 a month
196th of 365, with limits
$1,700 a month
Always-on AI employees
50 machines 24/7 · 2 vCPU / 4 GB
An assistant that runs all month fits a Zeabur server well, and costs less there. You set it up: install the agent over SSH, plus a desktop and browser if it needs them. boat has them installed.
6th cheapest of 361
$657 a month
157th of 361, with limits
$200 a month
Code interpreter bursts
100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month
Neither fits one-minute runs well. A Zeabur server bills a whole month. boat caps starts per minute, hour and day, and extra starts cost $0.03 to $0.07 each.
189th of 336, with limits
$2,000 a month
257th of 336, with limits
$300 a month
Sandboxes for your users
300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month
300 sandboxes for your users means 300 monthly servers on Zeabur, and Zeabur publishes no limit on how many one account can hold. boat allows 100 to 2,000 sandboxes at once by plan, plus start caps per plan.
3rd cheapest of 349
$209 a month
251st of 349, with limits
$1,200 a month
RL rollouts
1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month
39th of 349, with limits
$2,000 a month
249th of 349, with limits
$4,000 a month
Agent evals
250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month
Evals start many machines from the same state. boat forks a running sandbox in one call. Zeabur documents no whole-server snapshot or clone; you set it up: script the setup for each new server, and back up volumes and databases on the Dev plan or higher.
10th cheapest of 359
$200 a month
323rd of 359, with limits
$4,250 a month
CI runners
40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month
Zeabur runs build machines for the apps it deploys. For your own CI jobs, you set it up: install a runner on a server, which bills the full month.
2nd cheapest of 374
$86 a month
354th of 374, with limits
$2,280 a month
24/7 dev boxes
1 machine 24/7 · 4 vCPU / 8 GB
Both fit a 24/7 dev box, and Zeabur costs less for it. A Zeabur server has a fixed monthly price and a fixed IP. A boat sandbox saves its disk every minute; you turn auto-stop off (default 1 hour), and its IP changes at each resume.
19th cheapest of 371
$26 a month
149th of 371, with limits
$17 a month
Computer-use agents
50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month
boat streams a 60fps desktop with Chromium and computer-use tools. On Zeabur you set it up: install a desktop, a browser, a remote-display server and your own control loop on a plain Ubuntu server.
7th cheapest of 367
$200 a month
255th of 367, with limits
$850 a month
Browser-use agents
50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month
7th cheapest of 356
$200 a month
260th of 356, with limits
$200 a month
Scraping & bots (startup scale)
50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month
6th cheapest of 343
$90 a month
266th of 343, with limits
$150 a month
Mass scraping & bots (undetectable)
1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month
73rd of 327, with limits
$1,800 a month
112th of 327, with limits
$3,000 a month
Parallel red teaming
200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month
3rd cheapest of 351
$200 a month
307th of 351, with limits
$800 a month
GPU eval jobs
10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month
Not ranked: misses a requirementNot ranked: misses a requirement
macOS agents
5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month
Not ranked: misses a requirementNot ranked: misses a requirement

Key features

FeatureboatZeabur
What you getSandbox VMs for agents, created and deleted by API, CLI or SDKAn app platform: you deploy services, databases and templates onto a server
IsolationEvery sandbox is a full VM with its own kernelEach server is a whole VPS for you; your services run as containers on it
Getting a machineCreated from code when an agent needs one, billed per second; starts per minute, hour and day are capped by planBought in the dashboard or with Zeabur's tools (its Claude Code plugin can rent servers), billed per month and renewed automatically
StoppingA stopped sandbox has no compute charge. boat keeps its disk for at least 14 days; saved snapshots keep it longer.No documented pause that lowers the bill. A server bills monthly until you cancel it; unused time comes back as Zeabur credits.
Snapshots and forkDisk snapshot every minute and at each stop; fork a running sandbox (disk from the last 1-minute snapshot, no memory)Backups of volumes and databases (Dev plan and up, kept 7 days); the server disk persists for the life of the server; no whole-server snapshot or clone is documented
Docker and nested virtualizationDocker, root, systemd and /dev/kvm on every sandboxRoot over SSH on Ubuntu 24.04 or ZeaburOS; Docker Compose once Zeabur services are turned off; /dev/kvm not documented
Desktop and computer use60fps desktop and Chromium, computer-use and browser tools, embeddable streamYou set it up: install a desktop, a browser and a remote-display server on a plain Ubuntu server
Coding agentsSeven harnesses installed, including Claude Code, Codex and OpenCode, with a prompt, steer and events APIYou set it up: install each agent on a server over SSH. Zeabur's MCP server and Claude Code plugin let your own assistant deploy to Zeabur.
Keeping keys out of the machineOptional AI credentials proxy ($1 a day it is on, plus $1 per 1,000 requests); other secrets still enter the sandboxSecrets are environment variables injected into the service; you set it up: run your own key proxy to keep keys out
Reaching the machineHTTPS URLs for up to 50 ports, token-gated; public IPv6; the IP changes at each resumeFixed public IPv4 per server, HTTPS with automatic certificates on domains, inbound firewall rules by port and IP
Apps, databases, domains and emailSandboxes only; host your own services inside a VMOne-click templates, managed databases, domain purchase and transactional email
Your own serversRuns only on boat's machinesConnect any server with root SSH and 2 GB of memory, or an existing Kubernetes cluster
RegionsGermany, Finland and France; no US or Asia region; you cannot pick oneAsia, China, Europe, North America, South America and Oceania, across eight cloud providers; you pick the city
Compliance and transparencySOC 2 is in progress. We're open about where we stand: live status and data retention policy.No SOC 2 report published
Every feature Battleships tracks (43)
FeatureboatZeabur
Full VM (own kernel)YesYes
Docker insideYesYes
Nested virtualizationYesNot published
RootYesYes
SystemdYesYes
Desktop GUIYesNo documented way
BrowserYesNo documented way
Computer-use APIYesNo documented way
SSHYesYes
Persistent diskYesYes
Pause/resumeYesNo documented way
Automatic snapshotsYesNo documented way
Fork/cloneYesNo documented way
Live fork (no pause)YesNo documented way
Memory forkNo documented wayNo documented way
Custom image (Docker or snapshot)YesYes
≥24 h sessionsYesYes
Idle auto-stopNo documented wayNo documented way
Wake on requestNo documented wayNo documented way
HTTPS preview URLsYesYes
Raw TCP inboundYesNot published
Public IPv4No documented wayYes
Egress allowlistNo documented wayNo documented way
Secret proxyYesNo documented way
Secret proxy for any APIYesNo documented way
Preinstalled agentsYesNo documented way
Agent harness APIYesNo documented way
MCP serverNo documented wayYes
GPUNo documented wayNot published
WindowsNo documented wayNo documented way
Arm64No documented wayNot published
Self-hostingNo documented wayNo documented way
BYOCNo documented wayYes
Open sourceNo documented wayNo documented way
Scoped API keysYesNot published
Spend limitsYesYes
Usage per sandbox (API)YesNo documented way
Audit logsNo documented wayNot published
SSONo documented wayNo documented way
SOC 2No documented wayNo documented way
HIPAANo documented wayNo documented way
EU data residencyNo documented wayYes
Zero data retentionYesNot published

Performance

boat28.06 runs/s
ZeaburNot measured

Zeabur is not in the public benchmarks we use, so we cannot say which is faster. Its speed also depends on which provider's server you buy. In StarSling's run of 3 October 2026 (run 37073680030), boat did 28.06 runs/s on the V8 Web Tooling Benchmark at 4 vCPU / 8 GB, second of 15 providers. Every boat sandbox in this run ran on an AMD Ryzen 9 9950X host. boat also uses Ryzen 7 9700X hosts, and boat does not promise a CPU model.

V8 Web Tooling Benchmark at a 4 vCPU / 8 GB target, runs per second, higher is better. Source: StarSling's run of 3 October 2026 (run 37073680030), median of 5 sandboxes. Providers ran on different hardware in different regions. StarSling gives a shared rank when the run cannot separate two providers. Every boat sandbox in this run ran on an AMD Ryzen 9 9950X host. boat also places some sandboxes on AMD Ryzen 7 9700X hosts, which this run did not measure. Zeabur is not in the public benchmarks we use, so we cannot say which is faster.

Pricing in five scenarios

Each bill is priced from the providers' public price lists by the open Battleships engine, with the cheapest plan that runs the scenario. Start limits, per-start fees and plan minimums are in the price, boat's included.

Scenarioboat, a monthZeabur, a month
One coding agent, workdays
4 vCPU / 8 GB, 50 GB disk, 8 hours a day, 22 days a month, files kept between days
$20 · $20/mo$17 · Free · with limits
One agent that never stops
4 vCPU / 8 GB, 50 GB disk, running 24/7 all month
$26 · $20/mo$17 · Free · with limits
50 agents in parallel
50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside
$317 · $20/mo × 3 seats$850 · Free · with limits
Sandboxes for your users
2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once
$200 · $100/mo × 2 seats$1,200 · Free · with limits
1,000 short jobs at once
2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month
$8,800 · $2000/mo · with limits$4,000 · Free · with limits

boat: Plans cost $20, $100, $500 or $2,000 a month. The whole plan price comes back as VM time for that month, billed per second while a sandbox runs ($0.036 an hour at 4 vCPU / 8 GB). Included time does not carry over. Each plan caps starts (the $20 plan: 12 a minute, 60 an hour, 200 a day); extra starts cost $0.03 to $0.07 each. A stopped sandbox has no compute charge, but saved snapshots above the included count and the AI credentials proxy still charge. boat does not bill egress.

Zeabur: Each Zeabur server has a fixed monthly price that depends on the provider and city. It renews automatically until you cancel, busy or idle. A purchased server needs no Zeabur plan; plans add features such as backups. Our bills use the cheapest matching server in the catalog and no plan. If you cancel, Zeabur refunds the unused part as Zeabur credits (Tencent, Aliyun and Volcengine servers: one refund per account). Each server includes a traffic allowance; Zeabur publishes no price for traffic above it.

Limits in those scenarios

Zeabur is an app platform: you rent whole servers by the month instead of creating machines from code. So one agent on workdays costs the same as one running all month there, and for a single agent a Zeabur server (about $17) costs less than boat's $20 minimum plan. 50 agents or 300 sandboxes for your users mean 50 or 300 monthly servers, and Zeabur publishes no limit on how many an account can hold. On boat the limit is the number of sandboxes running at once (100 to 2,000 by plan) plus starts per minute, hour and day. boat never queues a start, so the 1,000-job burst hits the start caps and needs more starts per day than the $2,000 plan includes; it comes to $8,800 on boat against $4,000 for 1,000 Zeabur servers kept all month. boat keeps a stopped sandbox's disk for at least 14 days; a saved snapshot keeps it longer.

ScenarioboatZeabur
One coding agent, workdaysNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • egress price not published
  • monthly / always-on product: priced as 1 machines kept on and reused
One agent that never stopsNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • egress price not published
50 agents in parallelNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • egress price not published
  • monthly / always-on product: priced as 50 machines kept on and reused
Sandboxes for your usersNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • not confirmed: how many machines can run at once (you need 300; the typical published limit is 50)
  • egress price not published
1,000 short jobs at once
  • 2,400 starts/day included, paid above that (needs ~6,667)
  • 600 starts/hour included, paid above that (needs ~1,000)
  • starts above 90/min are paid, and starts above 270/min fail with 429
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • not confirmed: how many machines can run at once (you need 1,000; the typical published limit is 50)
  • egress price not published
boat's own limits
  • Plans start at $20 a month. The plan price comes back as VM time for that month. Unused time does not carry over. Source
  • Starts (create, fork, resume): the $20 plan includes 12 a minute, 60 an hour and 200 a day. The $2,000 plan includes 90 a minute, 600 an hour and 2,400 a day. Extra starts cost $0.03 to $0.07. A start fails with 429 above 3 times the included starts per minute, or 6 times per hour or per day. Source
  • A stopped sandbox has no compute charge. boat keeps its disk for at least 14 days. A saved snapshot keeps it longer; extra saved snapshots and the AI credentials proxy are paid. Source
  • Auto-stop defaults to 1 hour. You can turn it off on a paid plan. The trial allows 2 hours at most. Source
  • boat saves the disk every minute and at each stop. Memory, running programs and /tmp are not saved: a resume is like a reboot. A machine failure loses writes after the last good snapshot. Source
  • Machines run in Germany, Finland and France only. Snapshot copies can be stored outside the EU. Source
  • Public IPv6 and outbound IPv4. No dedicated IPv4. The IP changes at each resume. Source
  • Hosts are AMD Ryzen 7 9700X and Ryzen 9 9950X. vCPUs are shared. No GPU. You cannot boot your own Docker image. Source

Which one should you use?

Use Zeabur when

you host a web app and its databases, want a low fixed price for machines that run all month, need a region outside Europe, or want to use hardware you already own.

  • A full app platform. Useful if you are shipping a web app: Zeabur deploys services, databases, domains and email from one dashboard.
  • Lower cost for machines that run all month. Useful for always-on machines: a 4 vCPU / 8 GB Zeabur server costs about $17 a month.
  • Servers almost anywhere. Useful if your users are outside Europe: Zeabur sells servers in Asia, China, the Americas and Oceania.
  • Bring your own servers. Useful if you already own machines: Zeabur can manage any server with root SSH, or a Kubernetes cluster.

Use boat when

your agents need machines made from code: full VMs billed per second, with a disk snapshot every minute, fork while running, no compute charge while stopped, and a desktop and coding agents you drive by API.

  • Machines from code, billed per second. Agents come and go by the hour. On Zeabur you keep a pool of monthly servers and reuse them. boat's limits: starts per minute, hour and day are capped by plan, and memory and running programs are not saved at a stop.
  • Fork a running sandbox. Evals and parallel attempts start from the same state. On Zeabur you set it up: script the setup for each new server. On boat, a fork uses the last 1-minute snapshot, so writes from the last minute or so can be missing.
  • Coding agents and a desktop installed. On a Zeabur server you install each agent, desktop and browser, write the code that prompts and watches the agent, and put its API key in an environment variable that the agent can read.

Questions people ask

Is boat cheaper than Zeabur?

For one agent, no. A 4 vCPU / 8 GB Zeabur server costs about $17 a month, all month or on workdays, against $26 and $20 on boat. For fleets boat costs less: 50 agents in parallel come to $317 on boat against $850 on Zeabur, and sandboxes for your users to $200 against $1,200. For 1,000 short jobs at once Zeabur costs less, $4,000 against $8,800, and neither fits cleanly.

Can I run AI agents on Zeabur?

Yes, on a server you rent from it. You set it up: install the agent over SSH on plain Ubuntu, or deploy it as a container service. Zeabur's MCP server and Claude Code plugin let an assistant on your computer deploy apps to Zeabur. boat creates a VM per agent from code, with seven coding-agent harnesses installed and an API to prompt and steer them.

Is a Zeabur server a VM or a container?

The server is a whole VPS from a cloud provider, used only by you. With ZeaburOS, the services you deploy on it run as containers on K3s. With plain Ubuntu it is a normal VPS that you manage yourself. On boat, every sandbox is a full VM with its own kernel.

Can I pause a Zeabur server to stop paying?

Zeabur documents no pause that lowers the bill. A server bills monthly at a fixed rate and renews until you cancel it. On cancel, Zeabur refunds the unused part as Zeabur credits. On boat a stopped sandbox has no compute charge, and boat keeps its disk for at least 14 days.

Where does my data run?

boat sandboxes run on machines in Germany, Finland and France, and you cannot pick a region. Snapshot copies can be stored outside the EU. On Zeabur you choose a provider and city for each server, in Asia, China, Europe, North America, South America or Oceania.

Is boat SOC 2 compliant?

SOC 2 is in progress. We're open about where we stand: live status at status.ascii.dev and our data retention policy at docs.boat.dev/data-retention. Zeabur's security page says it is pursuing SOC 2 Type II certification.

Start a sandbox

Install the CLI, then run boat new.

7-day free trial

More comparisons