Cloudflare Sandbox vs Koyeb

Both run each sandbox in its own microVM and put it to sleep when idle. Cloudflare Sandbox lives inside Workers, tops out at 4 vCPU and lets you program every outbound request, while Koyeb runs sandboxes as services with GPUs, bigger machines and more ports to the outside.

The companies and products

Cloudflare Sandbox

Cloudflare Sandbox SDK runs untrusted or AI-generated code in Linux containers on Cloudflare's network, driven from a Cloudflare Worker.

Each sandbox is a Cloudflare Container backed by a Durable Object, and each container runs in its own Firecracker microVM with its own kernel and network. You build a linux/amd64 image, pick an instance type of up to 4 vCPU and 12 GiB, and reach the sandbox through your own Worker. It suits teams already on Workers who want short-lived sandboxes that sleep when idle; for code that only calls known methods, Cloudflare also offers Dynamic Workers, which run JavaScript in V8 isolates.

Koyeb

Koyeb is a serverless cloud for apps, inference and sandboxes, and in February 2026 it agreed to join Mistral AI.

A Koyeb sandbox is a Koyeb service: a microVM started from any Docker image and driven from Python or JavaScript SDKs that run commands, read files and expose a port. It can scale to zero when idle and wake on the next request, and it can run on CPU or GPU instance types. It is built for 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. “13th cheapest of 345” means 12 of the other 344 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 workloadCloudflare Sandbox: price rankKoyeb: price rankboat: price rank
Software factory
100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month
A software factory needs a disk that survives and room for Docker builds; Cloudflare resets the disk on sleep and Koyeb loses local disk on redeploy, while boat keeps the disk by default.
162nd of 367, with limits
$1,543 a month
54th of 367, with limits
$366 a month
3rd cheapest of 367
$634 a month
Always-on AI employees
50 machines 24/7 · 2 vCPU / 4 GB
207th of 366, with limits
$2,188 a month
135th of 366, with limits
$923 a month
5th cheapest of 366
$657 a month
Code interpreter bursts
100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month
Both suit short code runs: Cloudflare bills CPU only while busy and wakes on the next request, and Koyeb scales to zero after 5 idle minutes.
13th cheapest of 345
$144 a month
4th cheapest of 345
$66 a month
186th of 345, 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
Both fit sandboxes for your users; Cloudflare adds egress rules written as Worker code, Koyeb adds more regions, and boat gives each user a full VM with up to 50 HTTPS ports.
9th cheapest of 361
$731 a month
6th cheapest of 361
$495 a month
Cheapest of 361
$209 a month
RL rollouts
1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month
144th of 360, with limits
$9,428 a month
237th of 360, with limits
$1,631 a month
34th of 360, with limits
$2,000 a month
Agent evals
250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month
45th cheapest of 364
$818 a month
19th cheapest of 364
$429 a month
10th cheapest of 364
$200 a month
CI runners
40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month
Cloudflare caps disk at 20 GB by default and runs Docker with iptables off, Koyeb does not document /dev/kvm, and boat has full Docker plus /dev/kvm.
301st of 377, with limits
$704 a month
142nd of 377, with limits
$88 a month
4th cheapest of 377
$86 a month
24/7 dev boxes
1 machine 24/7 · 4 vCPU / 8 GB
A dev box should keep its files when you log off; Cloudflare starts from a fresh disk after sleep, Koyeb needs a volume, and boat stops at no cost with its disk kept.
70th cheapest of 368
$93 a month
217th of 368, with limits
$37 a month
18th cheapest of 368
$26 a month
Computer-use agents
50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month
Neither ships a desktop: Cloudflare removed its built-in one and Koyeb's SDK has none, while boat streams a 60fps desktop with Chromium.
166th of 369, with limits
$460 a month
81st of 369, with limits
$123 a month
6th cheapest of 369
$200 a month
Browser-use agents
50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month
210th of 367, with limits
$233 a month
172nd of 367, with limits
$123 a month
7th cheapest of 367
$200 a month
Scraping & bots (startup scale)
50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month
250th of 346, with limits
$286 a month
172nd of 346, with limits
$89 a month
6th cheapest of 346
$90 a month
Mass scraping & bots (undetectable)
1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month
169th of 338, with limits
$5,741 a month
65th of 338, with limits
$1,683 a month
71st of 338, with limits
$1,800 a month
Parallel red teaming
200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month
Red-team runs often need raw TCP and full root networking; Cloudflare takes no raw TCP from outside, Koyeb offers a TCP proxy to one port, and boat has raw TCP and nested VMs.
239th of 360, with limits
$325 a month
233rd of 360, with limits
$299 a month
2nd cheapest of 360
$200 a month
GPU eval jobs
10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month
Only Koyeb runs GPUs here, from RTX 4000 cards up to H100 and H200; Cloudflare Sandbox and boat are CPU only.
Can't run it3rd cheapest of 98
$628 a month
Can't run it
macOS agents
5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month
Can't run itCan't run itCan't run it

Key features

FeatureCloudflare SandboxKoyebboat
IsolationFirecracker microVM per container, own kernel and networkMicroVM per instanceFull VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts
Biggest machine4 vCPU / 12 GiB / 20 GB disk by default (can be raised)Up to 40 vCPU / 128 GB, and GPU types larger stillUp to 16 vCPU / 32 GB on the $100 plan and above
Files between sessionsFresh disk after sleep; snapshots (public beta, 30 days) or R2 backups when your code saves themLocal disk lost on redeploy; volumes (preview, 1 to 10 GB, two regions) or experimental SDK snapshotsDisk kept across stops and saved every minute; a stopped sandbox costs nothing
Fork and snapshotsNo fork; restore a saved snapshot or backup into a new sandboxSDK snapshots of the filesystem or with running processes, then new sandboxes from them (experimental)Fork a running sandbox without stopping it
Sleep and wakeStops after inactivity and wakes on the next request; CPU billed only while busyScales to zero after 5 idle minutes or more and wakes on request; Light Sleep wakes in about 200 msStop and start by API; TTL based, 1 hour by default
Docker insidedocker:dind with iptables and IP forwarding turned offStarts from any Docker image; Docker inside is supportedFull Docker, root and systemd
Nested virtualizationNot documentedNot documentedYes, /dev/kvm for Android emulators and VMs
Inbound trafficHTTP and WebSocket previews routed through your Worker; no raw TCP or public IPOne exposed HTTPS port at a time, plus an optional TCP proxy to one port; no dedicated IPHTTPS URLs for up to 50 ports, token-protected; raw TCP; own IPv6 or IPv4
Egress controlallowedHosts deny-by-default list plus outbound handlers you write in Worker codeBlock all outbound traffic, or allow only listed IPs and CIDRsOpen internet, no provider-side allowlist
Keeping keys out of the sandboxYour outbound handler adds the token for any HTTP or HTTPS host; you write the codeSecrets set as environment variables inside the serviceManaged proxy for model providers and any HTTPS API; switch a credential to another account
Desktop and browserBuilt-in desktop removed in SDK 0.10.2; bring your ownNo desktop or browser in the sandbox SDK60fps desktop with Chromium, computer-use API, embeddable stream
Coding agentsRun Claude Code, Codex, OpenCode or Pi in your image; OpenCode has an SDK helperTutorials that run agents through the command API; an MCP server lets agents manage KoyebClaude Code, Codex and five more preinstalled, with a prompt, steer and events API
GPUsNoRTX 4000 SFF Ada up to H100, H200 and multi-GPUNo
RegionsSeveral regions across Cloudflare's network, nearest free one pickedFrankfurt, Paris, Washington D.C., AWS US East, Singapore, TokyoEU only: Germany, Finland, France
Every feature Battleships tracks (43)
FeatureCloudflare SandboxKoyebboat
Full VM (own kernel)YesYesYes
Docker insideYesYesYes
Nested virtualizationNoNoYes
RootYesNot publishedYes
SystemdNot publishedNot publishedYes
Desktop GUINoNoYes
BrowserNoNoYes
Computer-use APINoNoYes
SSHYesNoYes
Persistent diskNoYesYes
Pause/resumeNoYesYes
Automatic snapshotsNoYesYes
Fork/cloneNoYesYes
Live fork (no pause)NoNot publishedYes
Memory forkNoYesNo
Custom image (Docker or snapshot)YesYesYes
≥24 h sessionsYesYesYes
Idle auto-stopYesYesNo
Wake on requestYesYesNo
HTTPS preview URLsYesYesYes
Raw TCP inboundNoYesYes
Public IPv4NoNoNo
Egress allowlistYesYesNo
Secret proxyYesNoYes
Secret proxy for any APIYesNoYes
Preinstalled agentsYesYesYes
Agent harness APIYesNoYes
MCP serverNoYesNo
GPUNoYesNo
WindowsNoNoNo
Arm64NoNoNo
Self-hostingNoYesNo
BYOCNoYesNo
Open sourceYesNoNo
Scoped API keysNot publishedNot publishedYes
Spend limitsNoNoYes
Usage per sandbox (API)Not publishedNoYes
Audit logsYesYesNo
SSOYesYesNo
SOC 2YesYesNo
HIPAAYesNoNo
EU data residencyYesYesYes
Zero data retentionNot publishedNot publishedYes

Performance

Cloudflare SandboxNot measured
KoyebNot measured
boat26.69 runs/s

Neither Cloudflare Sandbox nor Koyeb was part of the hpc-sandbox-benchmarks run, so there is no comparable CPU number for either. boat did 26.69 runs per second on that Node.js build-and-test loop at 4 vCPU / 8 GB. Koyeb publishes a CPU share of at least 0.25 dedicated vCPU per GB of memory, bursting to all cores, so its speed depends on how busy the host is.

Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Cloudflare Sandbox and Koyeb have 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.

ScenarioCloudflare Sandbox, a monthKoyeb, a monthboat, 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
$39 · Workers Paid · with limits$29 · Pro$20 · $20/mo
One agent that never stops
4 vCPU / 8 GB, 50 GB disk, running 24/7 all month
$148 · Workers Paid · 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
$1,660 · Workers Paid · with limits$475 · Pro$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
$476 · Workers Paid$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
$1,889 · Workers Paid · with limits$485 · Scale · with limits$2,000 · $2000/mo · with limits

Cloudflare Sandbox: You need Workers Paid ($5 a month minimum). CPU is billed only while in use, while memory and disk are billed on the instance size until the container sleeps, and there is no 4 vCPU / 8 GB size: 4 vCPU always comes with 12 GiB.

Koyeb: Koyeb bills each instance per second at a fixed price per type, and Deep Sleep costs nothing. Pro is $29 a month with $10 of usage included and Scale $299 with $100 included; the rest of the fee is not credit.

Koyeb: Plans start at $20 a month, and the whole fee is usage credit, billed per second while a sandbox runs. A stopped sandbox costs nothing and keeps its disk.

Limits in those scenarios

Cloudflare does not fit the scenarios that keep 50 GB of files between days, because its disk resets after sleep and tops out at 20 GB. Koyeb fits them, but its priced eco-xlarge shape has a 20 GiB disk with no published price for more. For 1,000 short jobs at once nobody fits as priced: Cloudflare's 1,500 vCPU account cap is short of the 2,000 needed, Koyeb's eco instances guarantee only half of each vCPU, and boat's $2,000 plan queues starts above 90 a minute.

ScenarioCloudflare SandboxKoyebboat
One coding agent, workdays
  • missing persistent disk
  • disk beyond 20 GiB: price unknown
None that bite
One agent that never stops
  • missing persistent disk
  • disk beyond 20 GiB: price unknown
None that bite
50 agents in parallel
  • missing persistent disk
None that biteNone that bite
Sandboxes for your usersNone that biteNone that biteNone that bite
1,000 short jobs at once
  • Workers Paid: max 1,500 vCPU running at once (needs 2,000)
  • burstable CPU: 50% of each vCPU guaranteed, throttled above that
  • max 2,400 starts/day (needs ~6,667)
  • max 600 starts/hour (needs ~1,000)
  • bursts above 90 starts/min are queued (a burst of 333 takes ~4 min to start)

Which one should you use?

Use Cloudflare Sandbox when

your app already runs on Workers and you want short sandboxes that sleep when idle, with egress rules you write as code.

  • Egress rules you can program. Cloudflare can block every host except the ones you list, and run your own Worker code on each outbound HTTP request.
  • CPU billed only while busy. Good for many short, spiky sessions: CPU is billed only while in use, and a sleeping sandbox wakes on the next request.
  • Sits next to your Workers. If your app already runs on Workers, sandboxes sit next to Durable Objects and R2, with less to wire up.

Use Koyeb when

your sandboxes need GPUs or bigger machines, should sleep and wake on a request, or must run in the US or Asia.

  • GPUs in the same platform. Useful when an agent or eval needs a GPU: Koyeb runs instances from RTX 4000 cards up to H100 and H200.
  • Bigger machines. Koyeb instance types go up to 40 vCPU and 128 GB, while a Cloudflare container tops out at 4 vCPU by default.
  • More ways in from outside. A Koyeb sandbox can expose an HTTPS port and a TCP proxy, where Cloudflare routes only HTTP and WebSocket through your Worker.

Where boat fits

Cloudflare Sandbox and Koyeb both shine at sandboxes that sleep between requests. If your agents instead work for hours on one machine, building, testing and running a browser, look at boat: a full VM that keeps its disk, runs nested VMs and a desktop, and comes with coding agents you drive from outside.

A disk that stays without extra code

On Cloudflare a sleeping sandbox wakes with a fresh disk, and on Koyeb local disk is lost on redeploy. Keeping a repo and installs means writing snapshot or backup code, or using Koyeb volumes capped at 10 GB in two regions.

Root, systemd, Docker and /dev/kvm

Cloudflare runs Docker with iptables turned off, and neither documents nested virtualization, so Android emulators and VM builds cannot start. boat gives full Docker, systemd and /dev/kvm on dedicated AMD Ryzen 9 9950X hosts.

Agents, desktop and ports built in

Neither ships a desktop, and Koyeb exposes one HTTPS port at a time with keys as environment variables the agent can read. boat has a 60fps desktop, 50 HTTPS ports and a credentials proxy that keeps real keys outside.

Use boat when your agents do real work for hours on a machine that keeps its disk: builds, full Docker, emulators, a desktop browser and preview servers, with agents and keys handled for you. Compare it one to one: boat vs Cloudflare Sandbox and boat vs Koyeb.

Questions people ask

Is Cloudflare Sandbox cheaper than Koyeb?

In the scenarios on this page, Koyeb costs less: $57 against $148 a month for one agent that never stops, and $329 against $476 for sandboxes for your users. Cloudflare's estimates for the always-on scenarios also do not fit, because its disk resets on sleep.

Which is faster, Cloudflare Sandbox or Koyeb?

There is no comparable measurement: neither ran the hpc-sandbox-benchmarks loop. Koyeb guarantees at least 0.25 dedicated vCPU per GB of memory, and Cloudflare caps a container at 4 vCPU by default. boat scored 26.69 runs per second on that benchmark.

Do Cloudflare Sandbox and Koyeb keep my files?

Not by default. A Cloudflare sandbox starts from a fresh disk after sleep unless your code saved a snapshot or R2 backup, and Koyeb loses local disk on redeploy unless you add a volume or an SDK snapshot. boat keeps the disk across stops and saves it every minute.

Can Cloudflare Sandbox or Koyeb run Docker?

Both start from Docker images. Cloudflare supports docker:dind with iptables and IP forwarding turned off, and neither documents /dev/kvm for emulators. boat runs full Docker with root, systemd and /dev/kvm.

Does either one have GPUs?

Koyeb does, from an RTX 4000 SFF Ada up to H100 and H200, billed per second. Cloudflare Sandbox and boat are CPU only.

Where does each one run?

Cloudflare places containers in several regions of its network and picks the nearest free one. Koyeb runs in Frankfurt, Paris, Washington D.C., Singapore, Tokyo and more. boat runs in the EU only, in Germany, Finland and France.

Try boat

Install the CLI, then run boat new.

7-day free trial

More comparisons