boat vs Alibaba Cloud Agent Sandbox
Both give each agent its own isolated Linux machine that speaks the E2B way of working. boat gives it a whole machine in the EU that runs as long as you want, with coding agents built in, while Alibaba's Agent Sandbox is strongest at short jobs that sleep and wake with their memory, in China, Asia and the US.
Where boat stands out
A whole machine with no daily cutoff
A boat sandbox is a full VM with root, Docker, systemd and /dev/kvm, up to 16 vCPU / 32 GB, with no session cap on paid plans and SSH when you want it. An Alibaba sandbox stops after 24 hours active at a time, and each one works on a 15 GB temporary disk that cannot be raised.
Coding agents you drive from outside
Claude Code, Codex and five other agents come preinstalled in every boat sandbox. You drive the agents inside from outside: prompt them, steer them and follow their events through one API. A credential can also switch to another account when one runs out of credits.
Your data stays in the EU
boat runs only in the EU, in Germany, Finland and France, with zero data retention. Alibaba's Agent Sandbox runs in mainland China, Hong Kong, Singapore and the US, with no EU region.
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
Alibaba Cloud Agent Sandbox
Alibaba Cloud Agent Sandbox (also sold as FC Agent Sandbox, and replacing the older AgentRun Sandbox) is Alibaba Cloud's sandbox service for AI agents, built on Function Compute.
It comes in two editions. Flash Edition is a serverless service you drive with the E2B SDK and CLI, with a second-generation runtime (MicroVM, in beta) that adds Docker, snapshots, forks and Windows. Container Edition runs sandboxes as pods in your own Alibaba Container Compute Service cluster. It is built for teams on Alibaba Cloud that want E2B-compatible sandboxes close to their other Alibaba services.
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 | Alibaba Cloud Agent Sandbox: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A coding agent on Alibaba's sandbox stops after 24 hours active and works on a 15 GB temporary disk, which a big build with its caches can fill; boat gives it a full VM with no session cap. | 3rd cheapest of 367 $634 a month | 6th cheapest of 367 $1,463 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB An assistant that lives on its machine all month has to pause and resume at least once a day on Alibaba; on boat it simply keeps running. | 5th cheapest of 366 $657 a month | 36th of 366, with limits $1,771 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Alibaba is built for this: short Code Interpreter runs billed per second, with light hibernation that wakes in about a millisecond. | 186th of 345, with limits $2,000 a month | 6th cheapest of 345 $90 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Alibaba's public URLs are open to anyone by default, and a token can only be turned on at creation; boat gives each user HTTPS URLs on up to 50 ports, token-protected by default. | Cheapest of 361 $209 a month | 8th cheapest of 361 $584 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month Alibaba's snapshots keep memory and fork one running state into many sandboxes, which suits rollouts that branch from the same point. | 34th of 360, with limits $2,000 a month | 46th of 360, with limits $2,732 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 | 21st cheapest of 364 $445 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI jobs that build Android apps or VM images need /dev/kvm, which boat exposes and Alibaba does not document. | 4th cheapest of 377 $86 a month | 23rd cheapest of 377 $283 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box needs SSH and no daily cutoff; Alibaba documents neither SSH access nor sessions past 24 hours. | 18th cheapest of 368 $26 a month | 243rd of 368, with limits $75 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Both fit; Alibaba also offers Windows desktops for computer use on its new runtime, and boat streams a Linux desktop at 60fps. | 6th cheapest of 369 $200 a month | 10th cheapest of 369 $400 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 | 59th of 367, with limits $198 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 | 10th cheapest of 346 $134 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 | 101st of 338, with limits $2,678 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 | 17th of 360, with limits $137 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
Every feature Battleships tracks (43)
| Feature | boat | Alibaba Cloud Agent Sandbox |
|---|---|---|
| Full VM (own kernel) | Yes | Yes |
| Docker inside | Yes | Yes |
| Nested virtualization | Yes | No |
| Root | Yes | Not published |
| Systemd | Yes | Not published |
| Desktop GUI | Yes | Yes |
| Browser | Yes | Yes |
| Computer-use API | Yes | Yes |
| SSH | Yes | No |
| Persistent disk | Yes | Yes |
| Pause/resume | Yes | Yes |
| Automatic snapshots | Yes | Yes |
| Fork/clone | Yes | Yes |
| Live fork (no pause) | Yes | Yes |
| Memory fork | No | Yes |
| Custom image (Docker or snapshot) | Yes | Yes |
| ≥24 h sessions | Yes | Yes |
| Idle auto-stop | No | Yes |
| Wake on request | No | Yes |
| HTTPS preview URLs | Yes | Yes |
| Raw TCP inbound | Yes | Not published |
| Public IPv4 | No | No |
| Egress allowlist | No | Yes |
| Secret proxy | Yes | Yes |
| Secret proxy for any API | Yes | Yes |
| Preinstalled agents | Yes | No |
| Agent harness API | Yes | No |
| MCP server | No | Yes |
| GPU | No | No |
| Windows | No | Yes |
| Arm64 | No | Not published |
| 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 | No |
| Audit logs | No | Not published |
| SSO | No | Yes |
| SOC 2 | No | Yes |
| HIPAA | No | Yes |
| EU data residency | Yes | No |
| Zero data retention | Yes | Not published |
Performance
There is no comparable measurement: Alibaba Cloud Agent Sandbox was not part of the hpc-sandbox-benchmarks run. boat did 26.69 runs per second on its Node.js build-and-test loop at 4 vCPU / 8 GB. Alibaba sells three compute classes, and only Performance promises a dedicated resource pool, so a fair test would need to pick one.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Alibaba Cloud Agent Sandbox 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 | Alibaba Cloud Agent Sandbox, 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 | $24 · Pay as you go |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $79 · Pay as you go · 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 | $916 · Pay as you go |
| 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 | $247 · Pay as you go |
| 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 | $546 · Pay as you go |
boat: Plans start at $20 a month, and the whole fee is usage credit. You pay per second while a sandbox runs, and a stopped sandbox costs nothing and keeps its disk.
Alibaba Cloud Agent Sandbox: Alibaba bills per second for the vCPU and memory you set, plus disk above a free 15 GiB, in three compute classes (Economy, Default and Performance), with a bill every hour. A sleeping sandbox pays for memory and disk, and deep hibernation and snapshots pay for memory plus disk size at the disk rate; mainland China regions cost less, and egress pricing is not published.
Limits in those scenarios
Alibaba stops a sandbox after 24 hours active at a time, so an agent that runs all month does not fit the always-on scenario without daily pauses. Alibaba allows 300 sandboxes per region by default, which covers the 300 user sandboxes, but its limits page does not say how to raise that for 1,000 jobs at once, and egress pricing is not published. boat has no session cap, but its start limits are lower: at 1,000 short jobs at once, the $2000 plan queues starts above 90 per minute and does not fit.
| Scenario | boat | Alibaba Cloud Agent Sandbox |
|---|---|---|
| One coding agent, workdays | None that bite |
|
| One agent that never stops | 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 Alibaba Cloud Agent Sandbox when
you already build on Alibaba Cloud, serve users in China or Asia, need Windows desktops, or run short jobs that sleep between requests and must wake instantly with their memory.
- Sleep and wake with memory. Useful for agents that sit idle between requests: a sleeping sandbox wakes in about a millisecond with its processes still running.
- Network controls and your own VPC. Teams can limit outbound traffic by IP, CIDR or domain and reach private databases through their own VPC.
- Windows, and regions in China, Asia and the US. Fits agents that need a Windows desktop, or that serve users in mainland China, Hong Kong, Singapore or the US.
Use boat when
your agents do real work on a machine for hours or days: installs, builds, tests, Docker, emulators and a browser, with coding agents built in and data that stays in the EU.
- A whole machine with no daily cutoff. A long agent task has to pause and resume every day on Alibaba, a big build can fill the disk, and Android emulators or VMs need /dev/kvm, which Alibaba does not document.
- Coding agents you drive from outside. Alibaba gives you a sandbox and frameworks for agents you write yourself. Running a coding agent there means installing it, keeping it updated and building your own way to send it work and read its output.
- Your data stays in the EU. Teams with European users or GDPR duties would have to send code and data outside the EU to use Alibaba's sandbox, and explain that to every customer who asks where it runs.
Questions people ask
Is boat cheaper than Alibaba Cloud Agent Sandbox?
boat costs less in four of the five scenarios on this page, and Alibaba does not fit the always-on one because of its 24-hour cap. For 1,000 short jobs at once Alibaba is cheaper, at $546 a month against $2,000, and boat does not fit, since its plans queue starts above 90 per minute.
How long can an Alibaba Agent Sandbox run?
Up to 86,400 seconds, which is 24 hours, in the active state at a time. You can pause it into deep hibernation and resume it from where it stopped. boat has no session cap on paid plans.
Can I move my E2B code to boat or to Alibaba?
Alibaba's Flash Edition works with the E2B SDK and CLI, with gaps listed in its docs, such as volumes and the metrics API, and SDK versions below 2.51.0. boat has its own API and CLI, so moving means rewriting the calls that create and drive sandboxes.
Can I run Docker or an Android emulator on Alibaba's sandbox?
Docker runs inside a sandbox on the second-generation runtime, which is in beta. Alibaba does not document /dev/kvm, which Android emulators and VMs need. boat runs Docker and exposes /dev/kvm in every sandbox.
Is Alibaba Cloud Agent Sandbox available in Europe?
Not today. It runs in four mainland China regions, Hong Kong, Singapore, Virginia and Silicon Valley. boat runs only in the EU, in Germany, Finland and France.
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