boat vs Paperspace
Paperspace, now part of DigitalOcean, rents Linux VMs with GPUs by the hour, and you reach them over SSH. boat gives each agent its own VM that you create, fork and drive from code, with a desktop, Chromium and coding agents already inside.
Where boat stands out
Built to be driven from code
On boat you create a sandbox, run commands, move files and fork it from an API, SDK or CLI. Paperspace's API and CLI create, start and stop machines, but its docs say you cannot use them to connect to a machine; you work inside over SSH.
Coding agents and a desktop inside
Every boat sandbox has Claude Code, Codex, OpenCode and other agents preinstalled, which you drive from outside with a prompt, steer and events API. It also has a 60fps desktop with Chromium for computer use. A Paperspace Linux machine is reached by SSH only, and desktop streaming is for Windows machines, which new users cannot create.
Snapshots every minute and fork while running
boat saves the disk every minute and on stop, a stopped sandbox costs nothing, and you can fork a running sandbox in one call. On Paperspace, auto-snapshots run at most every hour, the disk keeps billing while the machine is off, and a copy needs a custom template made from a machine that is turned off.
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
Paperspace
Paperspace is a GPU cloud for machine learning, owned by DigitalOcean since 2023.
Paperspace Machines are full Linux VMs with a persistent disk, a public IP and optional Nvidia GPUs from M4000 up to H100, billed by the hour. Next to Machines it sells Gradient Notebooks, Deployments and Workflows for training and serving models. It is built for people who train, test and run models, not as a sandbox API: its API manages machines, and you work inside them over SSH.
- Operating systems
- Linux; Windows templates closed to new users since 1 July 2024
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 | Paperspace: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month Paperspace gives you a real VM with Docker, but you set up each machine by hand over SSH and its API cannot run commands inside it, so a fleet of coding agents is a compromise. | 3rd cheapest of 367 $634 a month | 141st of 367, with limits $1,261 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB | 5th cheapest of 366 $657 a month | 179th of 366, with limits $1,560 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Neither is a great fit: boat's start limits slow big bursts, and Paperspace bills a full hour for each short run. | 186th of 345, with limits $2,000 a month | 179th of 345, with limits $1,451 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Paperspace has no API to run code or move files inside a machine, so handing sandboxes to your users means building that layer yourself; boat's VMs come with it. | Cheapest of 361 $209 a month | 124th of 361, with limits $1,243 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 | 193rd of 360, with limits $30,570 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 | 58th cheapest of 364 $1,241 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month Both fit: a Paperspace VM can host a CI runner, while boat starts and stops runners per job and bills per second. | 4th cheapest of 377 $86 a month | 101st cheapest of 377 $1,026 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB Both fit a dev box that never stops; Paperspace adds a disk and public IP fee on top of the hourly machine rate. | 18th cheapest of 368 $26 a month | 49th cheapest of 368 $63 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Paperspace streams a desktop only on Windows machines, which new users cannot create, so computer use is a compromise there; boat has a 60fps desktop with Chromium. | 6th cheapest of 369 $200 a month | 284th of 369, with limits $1,655 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 | 319th of 367, with limits $1,241 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 | 192nd of 346, with limits $112 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 | 92nd of 338, with limits $2,242 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 | 207th of 360, with limits $214 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month Only Paperspace fits: it rents Nvidia GPUs by the hour, and boat has no GPUs. | Can't run it | 30th cheapest of 98 $2,982 a month |
| 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 | Paperspace |
|---|---|---|
| Kind of product | Sandbox cloud for agents: VMs you create, fork and drive from code | GPU and CPU VM cloud for machine learning |
| Isolation | Every sandbox is a full VM with its own kernel | Every machine is a full VM with its own kernel |
| Docker and nested virtualization | Docker, root, systemd and /dev/kvm on every machine | Docker runs inside; nested virtualization is not documented |
| Working inside from code | API, SDK and CLI to run commands, move files and fork | The API and CLI manage machines but cannot connect to them; you use SSH |
| Coding agents | Claude Code, Codex, OpenCode and more preinstalled, with a prompt, steer and events API | None built in; ML templates such as ML-in-a-Box |
| Desktop and computer use | 60fps desktop and Chromium, computer use and browser control, embeddable stream | Desktop streaming only on Windows machines, which new users cannot create |
| Files after a stop | Snapshots every minute and on stop; a stopped sandbox costs nothing | Disk is kept and billed while off; auto-snapshots at most hourly, up to 10 kept |
| Copying a machine | Fork a running sandbox in one call | Custom templates; the machine must be turned off first, same region only |
| Reaching the machine | HTTPS URLs for up to 50 ports, token-protected; raw TCP; dedicated IPv6 or IPv4 | A public IPv4 on every Linux machine, any port; no managed HTTPS URLs; no IPv6 |
| Keeping keys out of the machine | AI credentials proxy for model providers and any HTTPS API you register | No credentials proxy; Gradient secrets are injected as environment variables |
| GPUs | None | Nvidia GPUs from M4000 to H100, up to 8 per machine |
| Largest CPU machine | 16 vCPU / 32 GB | C10: 32 vCPU / 244 GB |
| Regions | EU only: Germany, Finland, France | New York, Amsterdam and Canada |
| Billing step | Per second | Per hour |
| Compliance and transparency | SOC 2 is in progress. We're open about where we stand: live status and data retention policy. | SOC 2 |
Every feature Battleships tracks (43)
| Feature | boat | Paperspace |
|---|---|---|
| Full VM (own kernel) | Yes | Yes |
| Docker inside | Yes | Yes |
| Nested virtualization | Yes | Not published |
| Root | Yes | Not published |
| Systemd | Yes | Not published |
| Desktop GUI | Yes | No |
| 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 | No |
| Live fork (no pause) | Yes | No |
| Memory fork | No | No |
| Custom image (Docker or snapshot) | Yes | Yes |
| ≥24 h sessions | Yes | Yes |
| Idle auto-stop | No | Yes |
| Wake on request | No | No |
| HTTPS preview URLs | Yes | No |
| Raw TCP inbound | Yes | Not published |
| Public IPv4 | No | Yes |
| Egress allowlist | No | No |
| 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 | Yes |
| Windows | No | No |
| Arm64 | No | No |
| Self-hosting | No | No |
| BYOC | No | No |
| Open source | No | No |
| Scoped API keys | Yes | Not published |
| Spend limits | Yes | Yes |
| Usage per sandbox (API) | Yes | No |
| Audit logs | No | Not published |
| SSO | No | Yes |
| SOC 2 | No | Yes |
| HIPAA | No | Yes |
| EU data residency | Yes | Yes |
| Zero data retention | Yes | Not published |
Performance
There is no comparable measurement for Paperspace: it was not part of the public StarSling benchmark run. boat did 26.69 Node.js build-and-test runs per second at 4 vCPU / 8 GB in that run, on dedicated AMD Ryzen 9 9950X (Zen 5) hosts.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Paperspace 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 | Paperspace, 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 | $19 · Core usage |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $63 · Core usage |
| 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 | $804 · Core usage |
| 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 | $1,241 · Core usage |
| 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 | $8,274 · Core usage |
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.
Paperspace: Paperspace bills each machine per hour while it is on, at a fixed rate per size (C5, 4 vCPU / 8 GB, is $0.08 an hour). The boot disk (from $5 a month for 50 GB) and the public IP (up to $3 a month) keep billing until you delete the machine, even when it is off.
Limits in those scenarios
Paperspace bills by the hour, so a job that runs for five minutes is billed as a full hour; that is why sandboxes for your users and 1,000 short jobs cost so much more there. It has no start limits, so the 1,000-job burst fits on Paperspace, while on boat it is queued and needs more starts per day than the $2,000 plan allows. Paperspace does not publish a machine count limit; your team's limits sit on the billing tab and are raised by support ticket. Public traffic above 1 Gbps for more than 10 seconds is slowed to 200 Mbps.
| Scenario | boat | Paperspace |
|---|---|---|
| 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 | None that bite |
| Sandboxes for your users | None that bite | None that bite |
| 1,000 short jobs at once |
| None that bite |
Which one should you use?
Use Paperspace when
you need GPUs, very large CPU machines, or machines in the US or Canada that a person sets up and uses over SSH.
- GPUs by the hour. Useful if your agents train or test models: Paperspace rents Nvidia GPUs by the hour, from A4000 up to H100.
- Big machines and a public IP. Useful for heavy single jobs: CPU machines go up to 32 vCPU and 244 GB, each with its own public IPv4.
- Machines in the US and Canada. Useful if your users or data must stay in North America: Paperspace runs in New York and Canada as well as Amsterdam.
Use boat when
you run many agents at once and need each one in its own VM that code can create, fork and drive, with a desktop, coding agents and minute-by-minute snapshots built in.
- Built to be driven from code. An agent platform needs to start work in hundreds of machines without a person. On Paperspace you write and run your own SSH layer for that, and keep it working.
- Coding agents and a desktop inside. Without these, you install and update each agent yourself, and a computer-use agent on Paperspace needs a desktop and remote viewer that you set up on every machine.
- Snapshots every minute and fork while running. An agent that breaks its machine can lose up to an hour of work on Paperspace. Trying five ideas from one state means stopping the machine, making a template and starting five new machines.
Questions people ask
Is boat cheaper than Paperspace?
It depends on the shape. One coding agent on workdays costs about the same: $20 a month on boat and $19 on Paperspace. One agent that never stops comes to $26 versus $63, 50 agents in parallel $317 versus $804, and sandboxes for your users $200 versus $1,241. For 1,000 short jobs at once, Paperspace fits at $8,274 a month, while boat's $2,000 plan does not fit because of its start limits.
Is Paperspace a sandbox for AI agents?
No. Paperspace is a GPU cloud that rents full Linux VMs by the hour. You can run an agent on one, but its API only manages machines; you get inside them over SSH, and there is no built-in fork, desktop or agent API.
Does boat have GPUs like Paperspace?
No. boat has no GPUs; it runs Linux VMs on dedicated AMD Ryzen 9 9950X (Zen 5) hosts. If your agents train or test models on GPUs, Paperspace rents Nvidia cards up to H100.
Can I still get a Windows machine on Paperspace?
Only if you joined before 1 July 2024. Windows templates are closed to new Paperspace users, and desktop streaming only works on Windows machines. boat has no Windows guests either, but every sandbox has a Linux desktop with Chromium.
Where does my data run?
boat runs only in the EU (Germany, Finland and France), with EU data residency and a zero data retention option. Paperspace runs machines in New York, Amsterdam and Canada.
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. Paperspace's parent, DigitalOcean, states SOC 2 and SOC 3 and HIPAA eligibility on its trust page.
Start a sandbox
Install the CLI, then run boat new.
7-day free trial
More comparisons
- boat vs Verda
- boat vs Nebius
- boat vs Jarvislabs
- boat vs Cerebrium
- boat vs Runpod
- boat vs Crusoe
- boat vs Replicate
- boat vs SF Compute
- boat vs E2B
- boat vs Daytona
- boat vs Modal
- boat vs Vercel Sandbox
- Runpod vs Paperspace
- Nebius vs Paperspace
- Crusoe vs Paperspace
- Paperspace vs Jarvislabs
- Paperspace vs Hyperstack
- Paperspace vs Verda
- Paperspace vs Cerebrium
- Paperspace vs Baseten