bochka vs Alibaba Cloud AgentBay
Both give AI agents a cloud machine to work in. bochka gives each agent a full Linux VM that keeps its disk and can host services, while AgentBay offers ready-made browser, Windows desktop and Android phone sessions that are thrown away when they end.
Where bochka stands out
A whole Linux machine that keeps its disk
A bochka sandbox is a full VM with root, systemd, Docker and /dev/kvm for Android emulators and VMs. Its whole disk is saved every minute and when it stops, a stopped sandbox keeps it for free, and you can fork a running one. You can also SSH in like any server.
Hosting built in
Every bochka sandbox can serve HTTPS on up to 50 ports with a token by default, take raw TCP, and get its own IP, on every paid plan. Agents can run a web app, an API or a database and hand you a link.
Coding agents and keys built in
Claude Code, Codex and five other agents come preinstalled, and you prompt, steer and follow them from outside through one API. Real API keys stay outside the sandbox in a credentials proxy that works for any HTTPS API, and a credential can switch to another account when credits run out.
The companies and products
bochka
bochka.xyz 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 AgentBay
AgentBay is Alibaba Cloud's execution platform for AI agents: cloud browsers, desktops, Android phones and code environments behind one SDK.
You start a session from an image (Linux or Windows desktop, a browser, an Android phone, or a CodeSpace for code) and drive it through the SDK in Python, TypeScript, Go or Java, or through MCP. Each session runs in a VM with its own kernel, and it is released after a timeout unless you turn that off; only folders you save to a Context survive. It is built for teams that automate browsers, desktop apps and phones at scale on Alibaba Cloud.
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 | bochka: price rank | Alibaba Cloud AgentBay: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A software factory needs Docker, a disk that stays and ports you can open, which bochka has and AgentBay's CodeSpace does not document. | 3rd cheapest of 367 $634 a month | 214th of 367, with limits $2,541 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB An assistant that lives on its machine all month keeps its whole disk on bochka; on AgentBay only the folders you save to a Context survive a release. | 5th cheapest of 366 $657 a month | 14th cheapest of 366 $3,770 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month AgentBay's CodeSpace runs short Python and JavaScript jobs through a code API and releases the session when it is done, which fits this well. | 186th of 345, with limits $2,000 a month | 32nd cheapest of 345 $312 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Giving each user a sandbox with a web preview works on bochka's HTTPS ports on any plan; AgentBay links need Pro or Ultra and use ports 30100 to 30199. | Cheapest of 361 $209 a month | 254th of 361, with limits $1,170 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 | 89th of 360, with limits $5,081 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 | 149th of 364, with limits $943 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI that builds containers or Android emulators needs Docker and /dev/kvm, which bochka has. | 4th cheapest of 377 $86 a month | 75th cheapest of 377 $625 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box needs SSH, a disk that stays and services that keep running; AgentBay documents no SSH and releases idle sessions after 5 minutes by default. | 18th cheapest of 368 $26 a month | 274th of 368, with limits $145 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Both have desktops for computer use; AgentBay also has Windows desktops and its own computer-use agent, while bochka streams a Linux desktop at 60fps. | 6th cheapest of 369 $200 a month | 19th cheapest of 369 $863 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month AgentBay's browser has stealth, fingerprint and proxy options built in, while bochka gives the agent a full Chromium on a desktop it controls. | 7th cheapest of 367 $200 a month | 87th of 367, with limits $506 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 | 35th cheapest of 346 $391 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 | 159th of 338, with limits $4,985 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 | 66th of 360, with limits $385 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 | bochka | Alibaba Cloud AgentBay |
|---|---|---|
| Isolation | Full VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts | VM with its own guest kernel per sandbox |
| Operating systems | Linux (Ubuntu) | Linux and Windows desktops, Android phones, a browser image and CodeSpace |
| Docker inside | Yes, Docker preinstalled with root and systemd | Not documented; custom images are built from a Dockerfile outside the session |
| Nested virtualization | Yes, /dev/kvm for Android emulators and VMs | Not documented |
| What survives a stop | The whole disk, saved every minute and on stop; a stopped sandbox keeps it for free | Only folders you sync or mount from a Context; everything else is lost when the session is released |
| Pause and resume | Stop and start any time; resume boots the machine again with its disk | In-place pause that keeps session state, in beta and whitelist only |
| Fork | Fork a running sandbox with its disk | No fork or snapshot API; new sessions start from an image |
| Session length | No cap on paid plans; you set the time to live | Released after 5 idle minutes or 30 minutes by default; you can turn auto-release off |
| Inbound access | HTTPS URLs for up to 50 ports, token-protected by default; raw TCP; SSH; dedicated IPv4 or IPv6 | HTTPS or WebSocket links on Pro and Ultra, ports 30100 to 30199 by default; no public IP and no SSH |
| Outbound rules | Open internet, no provider-side allowlist | Security-group rules by IP, CIDR and port; domain rules on Windows sandboxes |
| Browser automation | Full Chromium on a 60fps desktop, driven by computer use or CDP | Managed browser over CDP with stealth, fingerprint, proxy options and slider CAPTCHA solving |
| Coding agents | Claude Code, Codex and five other agents preinstalled, with a prompt, steer and events API | Its own computer, browser and phone task agents; no coding agents preinstalled |
| Keeping API keys out of the sandbox | Credentials proxy for model providers and any HTTPS API; switch a credential to another account when credits run out | Keys are set as environment variables inside the session |
| Regions | EU only (Germany, Finland, France) | China, Singapore and the US |
| 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 | bochka | Alibaba Cloud AgentBay |
|---|---|---|
| Full VM (own kernel) | Yes | Yes |
| Docker inside | Yes | No |
| 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 | 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 | Yes |
| Raw TCP inbound | Yes | Not published |
| Public IPv4 | No | No |
| 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 | Yes |
| 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 | 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 for AgentBay. bochka did 26.69 runs per second on the StarSling hpc-sandbox-benchmarks Node.js build-and-test loop at 4 vCPU / 8 GB, on dedicated AMD Ryzen 9 9950X hosts. AgentBay has not been run on the same test, so we make no speed claim either way.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Alibaba Cloud AgentBay 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 | bochka, a month | Alibaba Cloud AgentBay, 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 | $31 · Basic |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $145 · Basic |
| 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 | $1,720 · Pro · 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 | $595 · Pro · 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 | $1,135 · Pro · with limits |
bochka: 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 AgentBay: AgentBay bills CPU and memory per second for as long as a session exists, idle or not, on top of the plan: Basic is free, Pro is $149 a month and Ultra is $225 a month, and the fee is not usage credit. The plan includes Context storage and 10 Mbps of basic bandwidth per instance; disk prices are not published.
Limits in those scenarios
AgentBay's free bandwidth is capped at a 10 Mbps peak per instance, which slows installs, image pulls and downloads in every scenario. Pro and Ultra allow 200 sessions at once, so 1,000 users or 1,000 short jobs need Ultra with extra slots, and Docker is not documented, so the 50-agent fleet does not fit. bochka has no session cap on paid plans, but its start limits are lower: for 1,000 short jobs at once, even the $2,000 plan queues starts.
| Scenario | bochka | Alibaba Cloud AgentBay |
|---|---|---|
| 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 AgentBay when
your agents drive Windows apps, Android phones or a stealth browser, or must run in China or Singapore.
- Windows desktops and Android phones. Useful when agents must drive Windows apps or Android apps: AgentBay has both as ready-made session images.
- A managed browser for scraping. Its browser image has stealth, fingerprint and proxy options and can solve slider CAPTCHAs.
- Network rules and Asia regions. Admins can limit traffic by IP, port or domain, and sessions run in China, Singapore or the US.
Use bochka when
your agents write and run code on Linux: Docker, builds, emulators and services they host, on a machine that keeps its disk between runs.
- A whole Linux machine that keeps its disk. On AgentBay, a released session is gone, and only folders you set up in a Context survive. An agent that installed packages, built images or left work outside those folders starts over next time.
- Hosting built in. AgentBay sessions have no public IP and no SSH. Its links need Pro or Ultra and open only ports 30100 to 30199 by default, so other ports wait on a whitelist request.
- Coding agents and keys built in. AgentBay's own agents drive desktops, browsers and phones, not your coding agent, and its docs put API keys in environment variables, where a prompt injection inside the session can read them.
Questions people ask
Is bochka cheaper than Alibaba Cloud AgentBay?
In four of the five scenarios on this page, yes. One agent that never stops is $26 a month on bochka against $145 on AgentBay, because AgentBay bills a running session even when it is idle. For 1,000 short jobs at once, neither fits as priced: bochka queues the starts and AgentBay Pro stops at 200 sessions.
Does AgentBay keep my files after a session ends?
Only what you save to a Context, which syncs or mounts chosen folders from object storage. Everything else is lost when the session is released. bochka keeps the whole disk, saved every minute and when the sandbox stops.
Can I run Docker in an AgentBay sandbox?
AgentBay's docs do not describe Docker inside a session; you build custom images from a Dockerfile with its CLI instead. bochka has Docker preinstalled with root, and /dev/kvm for Android emulators and VMs.
Can I SSH into an AgentBay sandbox or host a web app on it?
AgentBay documents no SSH and gives sandboxes no public IP. Pro and Ultra can open HTTPS links to ports 30100 to 30199. bochka gives SSH, HTTPS on up to 50 ports, raw TCP and its own IP.
Does AgentBay have an EU region?
No. AgentBay runs in Hangzhou, Singapore and Virginia. bochka runs only in the EU, in Germany, Finland and France.
Is bochka SOC 2 compliant?
SOC 2 is in progress. We're open about where we stand: live status at status.bochka.xyz and our data retention policy at docs.bochka.xyz/data-retention.
Start a sandbox
Install the CLI, then run bochka new.
Product preview