boat vs Kamatera
Kamatera is a cloud server host: you pick CPU, RAM and disk, get a KVM server with a public IP, and set it up yourself. boat starts sandbox VMs for agents from an API, saves them every minute, and ships a desktop and coding agents you drive from outside.
Where boat stands out
Nested virtualization and Docker ready
Every boat sandbox has /dev/kvm, Docker, systemd and root ready the moment it starts. Kamatera's own FAQ says it does not support nested virtualization on its cloud servers, so a VM inside the server is not an option there.
State saved every minute, and forks
boat snapshots a running sandbox every minute and again when it stops, and a stopped sandbox costs nothing. You can fork a running sandbox or save it as a named template. Kamatera can clone a server and take manual disk snapshots, but its automatic backup is a paid add-on that runs once a day.
Coding agents you drive from outside
boat ships Claude Code, Codex, OpenCode and other agents in every sandbox, and one API prompts them, steers them mid-task and streams their events. Its credentials proxy keeps real AI keys out of the sandbox, and a 60fps desktop with Chromium is ready for computer use. On Kamatera you install the agents, write the glue to run them over SSH, and keep keys on the server.
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.
- Isolation
- Full VM per sandbox, own kernel
- Regions
- EU: Germany, Finland, France
Kamatera
Kamatera is a cloud infrastructure provider, owned by The OMC Group (founded 1995), that rents cloud servers, block storage, private networks and firewalls in 21 data centers.
A Kamatera cloud server is a KVM virtual machine where you choose the CPU type, vCPU count, RAM and disk, with root access and a dedicated public IP. It is a general cloud host, not a sandbox product: you install Docker, agents, a desktop and TLS yourself, and drive servers from the console, a REST API, cloudcli or Terraform. It fits teams that want flexible servers in many regions, billed by the minute or by the month.
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. “3rd cheapest of 367” means 2 of the other 366 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 workload | boat: price rank | Kamatera: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A software factory needs Docker, /dev/kvm and saved state on many short-lived machines; boat has all three ready, while Kamatera has no nested virtualization and you build the rest yourself. | 3rd cheapest of 367 $634 a month | 109th of 367, with limits $969 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB Both can keep an assistant running for months on a real VM; boat adds snapshots every minute and a built-in agent API, while on Kamatera you install, back up and watch everything yourself. | 5th cheapest of 366 $657 a month | 145th of 366, with limits $1,001 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Many one-minute code runs fit Kamatera's per-minute billing better on paper, but no published limit says how many servers you can start at once, and boat's start limits make big bursts a compromise too. | 186th of 345, with limits $2,000 a month | 74th of 345, with limits $92 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Handing each of your users a fresh sandbox means 300 machines at once; boat's plans publish that count, while Kamatera does not publish how many servers one account can run. | Cheapest of 361 $209 a month | 48th of 361, with limits $345 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 34th of 360, with limits $2,000 a month | 32nd of 360, with limits $1,827 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month | 10th cheapest of 364 $200 a month | 102nd of 364, with limits $288 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month Both run CI jobs on full VMs billed for the time they run; boat starts them from an API with Docker ready, and Kamatera runners are servers you image and manage. | 4th cheapest of 377 $86 a month | 36th cheapest of 377 $336 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB Both fit a 24/7 dev box well: Kamatera gives a 4 vCPU / 8 GB server with a static IP, and boat gives a VM saved every minute that costs nothing while stopped. | 18th cheapest of 368 $26 a month | 29th cheapest of 368 $41 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 for computer use; a Kamatera server can run a desktop image, but the screen streaming and the agent loop are yours to set up. | 6th cheapest of 369 $200 a month | 130th of 369, with limits $295 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month | 7th cheapest of 367 $200 a month | 187th of 367, with limits $154 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month | 6th cheapest of 346 $90 a month | 164th of 346, with limits $74 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month | 71st of 338, with limits $1,800 a month | 54th of 338, with limits $1,427 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 2nd cheapest of 360 $200 a month | 125th of 360, with limits $72 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month | Can't run it | Can't run it |
| macOS agents 5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month | Can't run it | Can't run it |
Key features
| Feature | boat | Kamatera |
|---|---|---|
| What it is | Sandbox VMs for agents, started, stopped and forked from an API | Cloud servers you size, create and manage yourself |
| Isolation | Full VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts | KVM VM on Intel Xeon; shared threads on Type A, reserved threads or cores on B, D and T |
| Nested virtualization | /dev/kvm on every machine | Not supported on any cloud server |
| Docker, root and systemd | Ready on every machine | Full root on the server; you install Docker yourself |
| Saved state | Snapshots every minute and on stop; a stopped sandbox costs nothing | Manual disk snapshots, or a paid daily backup kept 14 days; disk and IP billed while powered off |
| Fork and templates | Fork a running sandbox; named snapshots as templates, shareable by code | Clone a server, also while it runs; disk images as templates (no memory kept) |
| Desktop and computer use | 60fps desktop and Chromium, computer use and browser control, embeddable stream | Desktop OS images and a separate Cloud Desktop product; no computer-use API |
| Coding agents | Claude Code, Codex, OpenCode and more preinstalled, with a prompt, steer and events API | None built in; you install agent frameworks yourself |
| Keeping keys out of the machine | AI credentials proxy for model providers and any HTTPS API you register | None; keys go in environment variables on the server |
| Reaching the machine | HTTPS URLs with TLS for up to 50 ports, token-protected; raw TCP; dedicated IPv6 or IPv4 | Dedicated datacenter IPv4, all ports; domains and TLS are yours to set up |
| Network rules | Open internet; no provider-side allowlist | Console firewall with inbound and outbound rules by IP, port and protocol; private VLANs |
| Machine sizes | Up to 16 vCPU / 32 GB on paid plans | Up to 32 vCPUs / 128 GB on shared Type A; up to 128 vCPUs / 512 GB on reserved Types B and T; CPU and RAM change on an existing server |
| Windows | No Windows guests; Wine for Windows apps | Windows Server and Windows Desktop images |
| Regions | EU only: Germany, Finland, France | 21 data centers in the US, Canada, Europe, Israel, Asia and Australia |
| Compliance and transparency | SOC 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)
| Feature | boat | Kamatera |
|---|---|---|
| Full VM (own kernel) | Yes | Yes |
| Docker inside | Yes | Yes |
| Nested virtualization | Yes | No |
| Root | Yes | Yes |
| Systemd | Yes | Yes |
| Desktop GUI | Yes | Yes |
| Browser | Yes | No |
| Computer-use API | Yes | No |
| SSH | Yes | Yes |
| Persistent disk | Yes | Yes |
| Pause/resume | Yes | Yes |
| Automatic snapshots | Yes | No |
| Fork/clone | Yes | Yes |
| Live fork (no pause) | Yes | Yes |
| Memory fork | No | No |
| Custom image (Docker or snapshot) | Yes | Yes |
| ≥24 h sessions | Yes | Yes |
| Idle auto-stop | No | No |
| Wake on request | No | No |
| HTTPS preview URLs | Yes | No |
| Raw TCP inbound | Yes | Yes |
| Public IPv4 | No | Yes |
| Egress allowlist | No | Yes |
| Secret proxy | Yes | No |
| Secret proxy for any API | Yes | No |
| Preinstalled agents | Yes | No |
| Agent harness API | Yes | No |
| MCP server | No | No |
| GPU | No | No |
| Windows | No | Yes |
| Arm64 | No | No |
| Self-hosting | No | No |
| BYOC | No | No |
| Open source | No | No |
| Scoped API keys | Yes | Not published |
| Spend limits | Yes | Not published |
| Usage per sandbox (API) | Yes | Yes |
| Audit logs | No | Not published |
| SSO | No | No |
| SOC 2 | No | No |
| HIPAA | No | No |
| EU data residency | Yes | Yes |
| Zero data retention | Yes | Not published |
Performance
Kamatera is not in the hpc-sandbox-benchmarks run, so there is no comparable CPU measurement for it. boat was measured on that Node.js build-and-test loop at 4 vCPU / 8 GB, and the table shows its score. Kamatera says it uses Intel Xeon Platinum and Ice Lake processors at 2.6 GHz or more; we don't guess how its CPU types would score.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Kamatera has no published run in this suite yet, so there is no like-for-like number to compare.
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.
| Scenario | boat, a month | Kamatera, 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 | $10 · Usage account |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $41 · Usage account |
| 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 | $484 · Usage account |
| 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 | $154 · Usage account · with limits |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $2,000 · $2000/mo · with limits | $366 · Usage account · with limits |
boat: You pick a monthly plan from $20, and all of it becomes usage credit spent per second on running sandboxes. A stopped sandbox costs nothing and keeps its disk.
Kamatera: Kamatera bills hourly servers by the minute for the CPU, RAM and disk you choose, plus $0.01 per GB of traffic after the trial, or rents a monthly plan with traffic included. A powered-off hourly server still pays a small hourly charge for its disk and IP. The prices below use the shared Type A CPU; reserved Types B and D cost more.
Limits in those scenarios
Kamatera publishes no limit on how many servers one account can run, so 50 agents look likely to fit, but 300 user sandboxes and 1,000 short jobs at once are not confirmed. The prices use Type A, which shares CPU threads with no guarantee, so heavy, sustained CPU use can be slowed by other tenants. On boat, concurrency goes from 100 to 2,000 by plan, but start-rate limits mean even the top plan can't launch 1,000 short jobs at once without queueing.
| Scenario | boat | Kamatera |
|---|---|---|
| One coding agent, workdays | None that bite | None that bite |
| One agent that never stops | None that bite | None that bite |
| 50 agents in parallel | None that bite |
|
| Sandboxes for your users | None that bite |
|
| 1,000 short jobs at once |
|
|
Which one should you use?
Use Kamatera when
you need servers outside Europe, Windows, very large machines, or firewall rules and private networks you control.
- Regions around the world. Useful if your users are far from Europe: Kamatera has data centers in the US, Canada, Israel, Asia and Australia.
- Windows servers and desktops. Useful if your agents must run real Windows software: Kamatera offers Windows Server and Windows Desktop images.
- Big machines and network control. Useful for very large servers or locked-down networks: up to 128 reserved vCPUs, firewall rules both ways and private VLANs.
Use boat when
agents build, test and run Docker or emulators on machines that start from an API, save themselves every minute, fork, and come with a desktop and coding agents you drive from outside.
- Nested virtualization and Docker ready. Android emulators, Firecracker, test VMs and some build tools need /dev/kvm. On Kamatera there is no way to turn it on, so those jobs need a different host.
- State saved every minute, and forks. Agents sometimes break an install or delete files. With a daily backup you can lose up to a day of work, or you must remember to take a manual snapshot before each risky step.
- Coding agents you drive from outside. On a plain cloud server the agent's API keys sit in environment variables the agent can read and leak. You also build and maintain the code that starts, prompts and watches each agent.
Questions people ask
Is boat cheaper than Kamatera?
It depends on the shape of the work. One coding agent on workdays is $10 a month on Kamatera against $20 on boat. One agent that never stops is $26 on boat against $41, and 50 agents in parallel come to $317 on boat against $484. Kamatera's prices use its shared Type A CPU.
Does Kamatera support nested virtualization?
No. Kamatera's FAQ says it does not support nested virtualization on its cloud servers. Every boat sandbox has /dev/kvm, so Android emulators, Firecracker and test VMs run inside it.
Can I run Claude Code on a Kamatera server?
Yes, you get root and can install it yourself, then run it over SSH with your API key on the server. boat preinstalls Claude Code, Codex, OpenCode and others, drives them through an API, and keeps real keys out of the sandbox with its credentials proxy.
Does Kamatera charge for a powered-off server?
Yes, a small hourly charge for the disk space and IP address while an hourly server is powered off. On boat a stopped sandbox costs nothing and keeps its disk.
Where do boat and Kamatera run?
boat runs only in the EU, in Germany, Finland and France, with EU data residency and a zero data retention option. Kamatera has 21 data centers in the US, Canada, Europe, Israel, Asia and Australia.
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.
Start a sandbox
Install the CLI, then run boat new.
7-day free trial