Railway vs Koyeb
Both are app platforms that now sell sandboxes as VMs. Railway destroys an idle sandbox after its idle timeout, and files survive in checkpoints you take. Railway's Cloud Agents keep their disk, but they are in Priority Boarding. A Koyeb sandbox can sleep, wake on the next request and run on a GPU, and SDK snapshots keep its files.
The companies and products
Railway
Railway is a cloud platform for deploying web apps, servers and databases. It also sells Sandboxes and Cloud Agents.
Railway's main product is app hosting: you deploy services and databases as containers on its own hardware. In 2026 it added Sandboxes: isolated Linux VMs that you create from the CLI or a TypeScript SDK, run commands in and destroy. It also added Cloud Agents: a personal VM for a coding agent that keeps its disk and sleeps or wakes. Cloud Agents are in Priority Boarding, and breaking changes may occur. Railway suits teams that already host on Railway and want sandboxes next to their apps.
Koyeb
Koyeb is a serverless cloud for apps, inference and sandboxes. In February 2026 it agreed to join Mistral AI.
A Koyeb sandbox is a Koyeb service: a microVM that starts from any Docker image. You drive it from Python or JavaScript SDKs or the Koyeb CLI: run commands, read files, expose a port and take snapshots. The default image comes with four coding agents. A sandbox can scale to zero when idle and wake on the next request, on CPU or GPU instance types. It suits teams that run AI agents or untrusted code next to apps they already host on Koyeb.
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. “28th cheapest of 365” means 27 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 workload | Railway: price rank | Koyeb: price rank | boat: price rank |
|---|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month On both, you set up persistence. Railway keeps files in checkpoints you take, or in a Cloud Agent (Priority Boarding). Koyeb keeps them in SDK snapshots or a volume (preview, 1 to 10 GB). boat saves the disk every minute and at each stop with no code. | 28th cheapest of 365 $1,481 a month | 92nd of 365, with limits $366 a month | 5th cheapest of 365 $634 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB Railway fits an always-on assistant if you set the idle timeout to 0, and memory then bills the whole time. On Koyeb you keep the sandbox awake, or take an SDK snapshot before Deep Sleep discards the VM. On boat you turn auto-stop off (default 1 hour). | 21st cheapest of 361 $3,977 a month | 194th of 361, with limits $923 a month | 6th cheapest of 361 $657 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Both fit short code runs. Koyeb has small instances from $0.0036 an hour and scale to zero. Railway bills for CPU used. boat's $20 plan includes 12 starts per minute and 200 per day. | 42nd cheapest of 336 $413 a month | 6th cheapest of 336 $66 a month | 189th of 336, with limits $2,000 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Koyeb fits sandboxes for your users and wakes them on a request. Railway allows 100 sandboxes per environment on Pro, so 300 need several environments or an Enterprise limit. boat allows 300 at once from the $100 plan. | 46th of 349, with limits $290 a month | 8th cheapest of 349 $495 a month | 3rd cheapest of 349 $209 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 84th of 349, with limits $4,450 a month | 233rd of 349, with limits $1,631 a month | 39th of 349, with limits $2,000 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month Railway can fork a running sandbox but caps how many run per environment. Koyeb fits. boat forks a running sandbox from its last 1-minute snapshot and limits starts per plan. | 113th of 359, with limits $364 a month | 27th cheapest of 359 $429 a month | 10th cheapest of 359 $200 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI that builds Android apps or VM images needs /dev/kvm. boat documents it on every sandbox. Neither Railway nor Koyeb documents nested virtualization. | 116th cheapest of 374 $1,210 a month | 143rd of 374, with limits $88 a month | 2nd cheapest of 374 $86 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB | 69th cheapest of 371 $89 a month | 232nd of 371, with limits $37 a month | 19th cheapest of 371 $26 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Neither Railway nor Koyeb includes a desktop. On Railway you add Xvfb and noVNC with the template builder and publish the port. On Koyeb you add a desktop and VNC to your own Docker image. boat includes a 60 fps desktop with Chrome. | 47th cheapest of 367 $407 a month | 128th of 367, with limits $123 a month | 7th cheapest of 367 $200 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month | 53rd of 356, with limits $93 a month | 237th of 356, with limits $123 a month | 7th cheapest of 356 $200 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month | 65th cheapest of 343 $570 a month | 242nd of 343, with limits $89 a month | 6th cheapest of 343 $90 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month | 111th of 327, with limits $2,988 a month | 67th of 327, with limits $1,683 a month | 73rd of 327, with limits $1,800 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 71st of 351, with limits $574 a month | 261st of 351, with limits $299 a month | 3rd cheapest of 351 $200 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month Koyeb runs GPU instances from an RTX 4000 SFF Ada up to H100, H200 and B200. Railway sandboxes and boat have no GPU. | Not ranked: misses a requirement | 3rd cheapest of 102 $628 a month | Not 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 requirement | Not ranked: misses a requirement | Not ranked: misses a requirement |
Key features
| Feature | Railway | Koyeb | boat |
|---|---|---|---|
| Isolation | Sandboxes and Cloud Agents are isolated Linux VMs; app services are containers | Cloud Hypervisor microVM per instance | Full VM with its own kernel, on physical hosts dedicated to boat; vCPUs are shared |
| When the sandbox goes idle | Destroyed after the idle timeout (30 minutes by default, or 0 for never); Cloud Agents sleep and keep their disk | Scales to zero after at least 5 idle minutes and wakes on the next request; Light Sleep (preview) wakes in about 200 ms; not on eco instances | Auto-stop after 1 hour by default, a timer that does not watch for idle; no compute charge while stopped |
| Files that survive | Disk checkpoints you take before the sandbox is destroyed; they boot new sandboxes | Local disk is lost on redeploy and in Deep Sleep; SDK snapshots save the filesystem for new sandboxes; volumes (preview, 1 to 10 GB, two regions) persist | Disk saved every minute and at each stop with no code; kept at least 14 days after a stop; memory and running programs are not saved |
| Fork and snapshots | Fork a running sandbox's filesystem; boot new ones from named checkpoints (disk only) | SDK snapshots of the filesystem, or with running processes, then new sandboxes from them | Fork a running sandbox from its last 1-minute snapshot; saved snapshots start new ones (disk only) |
| Start from your own image | Templates built from ordered steps; no custom Docker or OCI image for sandboxes | Any Docker image | No; boat's Ubuntu 24.04 image or a saved snapshot |
| Docker and nested virtualization | Root shell on an Ubuntu VM; Docker and /dev/kvm are not documented, so you install and verify them yourself | Sandboxes start from Docker images; Docker inside a sandbox and nested virtualization are not documented | Docker, root, systemd and /dev/kvm documented on every sandbox |
| Reaching the sandbox | Up to 10 public HTTPS domains, set only at creation, with no built-in auth; SSH and port forwarding | One exposed HTTPS port at a time, plus an optional TCP proxy to one port (preview) | Up to 50 HTTPS URLs, token-gated by default; SSH; public IPv6 for raw ports; no dedicated IPv4, and the IP changes at each resume |
| Outbound rules | No egress allowlist | Block all outbound traffic, or allow only listed IPs and CIDRs, from the SDK | No API setting; you add firewall rules inside the sandbox |
| Coding agents | Claude Code, Codex, OpenCode and Pi in the image, run through exec; Cloud Agents keep terminal sessions you reattach to | Default image installs Mistral Vibe, Codex, Gemini and OpenCode, run through the command API | Seven harnesses installed, with a prompt, steer and events API |
| Keeping API keys out of the sandbox | Secrets enter as environment variables, or you sign in and save it in a checkpoint | Secrets are environment variables inside the service | Optional paid proxy ($1 a day when on, plus $1 per 1,000 requests) keeps real keys of supported model providers out; other secrets still enter |
| Desktop and computer use | No desktop included; you can add Xvfb and noVNC with the template builder and publish the port on a sandbox domain | No desktop or browser preinstalled; you can add them to your own Docker image | Built-in desktop and Chrome stream at 1920x1080, 60 fps, with computer-use tools |
| GPUs | No GPU for sandboxes | RTX 4000 SFF Ada up to H100, H200 and B200, and multi-GPU; some types on request | No GPU on any size |
| Apps and databases next to the sandbox | Postgres, Redis and services in the same project; a sandbox can join their private network | Apps and inference run as Koyeb services on the same platform | No hosted databases or app services; host your own inside a sandbox |
| Biggest machine | 8 vCPU / 8 GB on Hobby; on Pro up to 32 vCPU / 32 GB per the docs (24 vCPU / 24 GB per the pricing page) | Instance types up to 40 vCPU / 128 GB, and GPU types larger still | 8 vCPU / 16 GB / 111 GB disk on the $20 plan; 16 vCPU / 32 GB / 251 GB disk from the $100 plan |
| Regions | US West, US East, Amsterdam and Singapore | Frankfurt, Paris, Washington D.C., AWS US East, Singapore, Tokyo, plus San Francisco in preview | Germany, Finland and France; you cannot pick a region; no US or Asia region |
Every feature Battleships tracks (43)
| Feature | Railway | Koyeb | boat |
|---|---|---|---|
| Full VM (own kernel) | Yes | Yes | Yes |
| Docker inside | Not published | Not published | Yes |
| Nested virtualization | Not published | No documented way | Yes |
| Root | Yes | Not published | Yes |
| Systemd | Not published | Not published | Yes |
| Desktop GUI | With your own setup | No documented way | Yes |
| Browser | Yes | No documented way | Yes |
| Computer-use API | No documented way | No documented way | Yes |
| SSH | Yes | No documented way | Yes |
| Persistent disk | Yes | With your own setup | Yes |
| Pause/resume | Yes | Yes | Yes |
| Automatic snapshots | No documented way | Yes | Yes |
| Fork/clone | Yes | Yes | Yes |
| Live fork (no pause) | Yes | Not published | Yes |
| Memory fork | No documented way | Yes | No documented way |
| Custom image (Docker or snapshot) | Yes | Yes | Yes |
| ≥24 h sessions | Yes | Yes | Yes |
| Idle auto-stop | Yes | Yes | No documented way |
| Wake on request | No documented way | Yes | No documented way |
| HTTPS preview URLs | Yes | Yes | Yes |
| Raw TCP inbound | No documented way | Yes | Yes |
| Public IPv4 | No documented way | No documented way | No documented way |
| Egress allowlist | No documented way | Yes | No documented way |
| Secret proxy | No documented way | No documented way | Yes |
| Secret proxy for any API | No documented way | No documented way | Yes |
| Preinstalled agents | Yes | Yes | Yes |
| Agent harness API | Yes | No documented way | Yes |
| MCP server | Yes | Yes | No documented way |
| GPU | No documented way | Yes | No documented way |
| Windows | No documented way | No documented way | No documented way |
| Arm64 | No documented way | No documented way | No documented way |
| Self-hosting | No documented way | Yes | No documented way |
| BYOC | Yes | Yes | No documented way |
| Open source | No documented way | No documented way | No documented way |
| Scoped API keys | Not published | Not published | Yes |
| Spend limits | Yes | No documented way | Yes |
| Usage per sandbox (API) | No documented way | No documented way | Yes |
| Audit logs | Yes | Yes | No documented way |
| SSO | Yes | Yes | No documented way |
| SOC 2 | Yes | Yes | No documented way |
| HIPAA | Yes | No documented way | No documented way |
| EU data residency | Yes | Yes | No documented way |
| Zero data retention | Not published | Not published | Yes |
Performance
Railway and Koyeb are not in the public benchmarks we use, so we cannot say which is faster. 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. Every boat sandbox in that run ran on an AMD Ryzen 9 9950X host. boat also places sandboxes on other hosts and does not promise a CPU model. Koyeb publishes a CPU share of at least 0.25 dedicated vCPU per GB of memory, bursting to all cores.
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. Railway and Koyeb are 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.
| Scenario | Railway, a month | Koyeb, a month | boat, 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 | $16 · Hobby · with limits | $29 · Pro | $20 · $20/mo |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $70 · Hobby · with limits | $57 · Pro | $26 · $20/mo |
| 50 agents in parallel 50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside | $739 · Hobby · with limits | $475 · Pro · with limits | $317 · $20/mo × 3 seats |
| Sandboxes for your users 2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once | $223 · Pro · with limits | $329 · Scale | $200 · $100/mo × 2 seats |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $781 · Pro · with limits | $485 · Scale · with limits | $8,800 · $2000/mo · with limits |
Railway: Railway bills sandboxes per second at about $50 per vCPU-month on CPU actually used and $50 per GB-month on memory in use, plus $0.05 per GB out. Memory keeps billing while a sandbox waits, until it is destroyed. Plans: Free ($0 with $1 of usage a month), Hobby ($5) and Pro ($20 per workspace); the Hobby and Pro fees count toward usage. Railway does not publish a sandbox disk price, so our bills leave disk out.
Koyeb: Koyeb bills each instance per second from the moment it starts, at a fixed price per instance type: eco-xlarge (4 vCPU / 8 GB, 20 GB disk) costs $0.0576 an hour. The idle period before sleep bills at the full rate, and Deep Sleep costs nothing. Pro costs $29 a month with $10 of usage included, and Scale $299 with $100 included; the rest of the fee is not credit.
Koyeb: Plans cost $20, $100, $500 or $2,000 a month, and the whole plan price comes back as VM time for that month. A default sandbox (4 vCPU / 8 GB) costs $0.036 an hour, billed per second while it runs, busy or idle. A stopped sandbox has no compute charge, but saved snapshots above the included count and the AI credentials proxy still charge. The $20 plan allows 100 sandboxes at once and 12 starts per minute, 60 per hour, 200 per day. Extra starts cost $0.03 to $0.07 each.
Limits in those scenarios
We mark every Railway scenario as not a clean fit because Railway publishes no sandbox disk price, so its bills count CPU and memory only. Railway caps sandboxes per environment, 50 on Hobby and 100 on Pro, so 300 user sandboxes or 1,000 jobs at once need several environments or an Enterprise limit. Koyeb's priced shape for 4 vCPU / 8 GB is eco-xlarge, with a 20 GB disk that takes no volume, so it does not keep a 50 GB disk. Koyeb does not document Docker inside a sandbox, so the 50-agent fleet does not fit as priced. At 1,000 short jobs at once, boat's $2,000 plan bills starts above its included 90 a minute, 600 an hour and 2,400 a day at $0.03 to $0.07 each, and refuses starts above 270 a minute with 429.
| Scenario | Railway | Koyeb | boat |
|---|---|---|---|
| One coding agent, workdays |
|
| None for this scenario |
| One agent that never stops |
|
| None for this scenario |
| 50 agents in parallel |
|
| None for this scenario |
| Sandboxes for your users |
| None for this scenario | None for this scenario |
| 1,000 short jobs at once |
|
|
|
Which one should you use?
Use Railway when
your app and databases already live on Railway and you want sandboxes on the same private network, billed for the CPU they use.
- Your apps and databases in one place. Useful if your app already runs on Railway: sandboxes can join the same private network as your Postgres and services.
- Bills for CPU actually used. Useful for agents that mostly wait: Railway meters sandbox CPU on what it uses. Memory keeps billing until the sandbox is destroyed.
- SSH and a forkable running sandbox. Useful when you want to shell in and branch work: Railway offers SSH and forks a running sandbox's filesystem.
Use Koyeb when
your sandboxes need GPUs, sleep most of the time and must wake on a request, start from your own Docker image, or need outbound traffic locked down.
- GPUs on the same platform. Useful when an agent or eval needs a GPU: Koyeb runs instances from RTX 4000 cards up to H100, H200 and B200.
- Sleep when idle, wake on request. Good for sandboxes that sit idle between requests: Koyeb scales to zero and wakes on the next one.
- Your own image and outbound rules. Koyeb starts a sandbox from any Docker image, can block or allowlist outbound traffic by IP, and runs in Europe, the US and Asia.
Where boat fits
Railway and Koyeb are app platforms first, and on both you set up persistence for a sandbox yourself. For agents that work for hours or days, boat gives each one a full VM that saves its disk with no code, with Docker, /dev/kvm, a desktop and coding agents you drive by API. boat runs only in Germany, Finland and France, and limits starts by plan.
A disk saved for you
A Railway sandbox is destroyed after 30 idle minutes by default, and only the checkpoints you took survive. On Koyeb, Deep Sleep discards the VM, so you take SDK snapshots or attach a preview volume. A boat sandbox saves its disk every minute and at each stop, with no compute charge while stopped. Memory and running programs are not saved.
Docker, /dev/kvm and a desktop, documented
Neither Railway nor Koyeb documents Docker inside a sandbox or /dev/kvm, and neither includes a desktop. You can install a desktop on both yourself. boat documents root, systemd, Docker and /dev/kvm, and includes a 60 fps desktop with Chrome. boat has no GPU.
Agents you drive, keys kept outside
On both, you run agents through exec or a command API and parse the output yourself, and API keys are environment variables the agent can read. boat adds a prompt, steer and events API. Its optional paid proxy keeps the real keys of supported model providers outside the sandbox.
Use boat when your agents do real work for hours on a full VM that saves its disk for you: builds, Docker, emulators, a desktop browser and preview servers, with agents you drive by API. boat runs only in the EU and limits starts by plan. Compare it one to one: boat vs Railway and boat vs Koyeb.
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
Questions people ask
Is Railway cheaper than Koyeb?
It depends on the job. Railway is lower for one agent on workdays ($16 against $29) and for sandboxes for your users ($223 against $329), but its numbers leave out disk. Koyeb is lower for one agent that never stops ($57 against $70), but its priced shape has a 20 GB disk. boat comes to $20, $200 and $26 for those three jobs.
Which is faster, Railway or Koyeb?
We cannot say: neither is in the public benchmarks we use. boat did 28.06 runs/s on the V8 Web Tooling Benchmark at 4 vCPU / 8 GB in StarSling's run of 3 October 2026, on an AMD Ryzen 9 9950X host. Koyeb guarantees at least 0.25 dedicated vCPU per GB of memory and bursts above that.
What happens to an idle sandbox on Railway and on Koyeb?
Railway destroys it once the idle timeout runs out, 30 minutes by default, so you take a checkpoint first to keep files. Koyeb scales it to zero and wakes it on the next request; Deep Sleep discards the VM, so you take an SDK snapshot first. A boat sandbox auto-stops after 1 hour by default, keeps its disk for at least 14 days, and has no compute charge while stopped.
Can Railway or Koyeb sandboxes run Docker and Android emulators?
Railway's sandbox docs do not mention Docker or /dev/kvm, so you install and verify them yourself. Koyeb starts sandboxes from Docker images but does not document Docker inside a sandbox or nested virtualization. boat documents Docker, root, systemd and /dev/kvm in every sandbox, but cannot boot your own Docker image.
Does either have GPUs?
Koyeb does, from an RTX 4000 SFF Ada up to H100, H200 and B200, billed per second. Railway sandboxes and boat have no GPU.
Is boat SOC 2 compliant?
SOC 2 is in progress, so boat has no SOC 2 report today. Live status is at status.ascii.dev and the data retention policy is at docs.boat.dev/data-retention. Railway's docs say it is SOC 2 Type II certified, and Koyeb lists ISO 27001 and SOC 2 on its Enterprise plan.
Try boat
Install the CLI, then run boat new.
7-day free trial