boat vs Modelence
Modelence is a platform that runs your Node.js backend, databases and scheduled jobs in production and watches them for you. boat gives each agent its own full Linux VM, with Docker, root, a disk saved every minute and coding agents you drive from outside. The two overlap only where an agent needs a machine to work in.
Where boat stands out
A whole machine, not one web app
Every boat sandbox is a full VM with its own kernel, root, systemd, Docker and /dev/kvm. An agent can install any language, run several services and boot an Android emulator. Modelence runs one Node.js web process per app, with scheduled jobs through its framework.
A disk that keeps the agent's work
A boat sandbox saves its disk every minute and at each stop, and keeps it for at least 14 days after a stop. You can fork a running sandbox from its last 1-minute snapshot to try two ideas from the same point. Writes from the last minute or so can be missing from a fork or after a machine failure. Modelence rebuilds its container from your repository on each deploy.
Coding agents you drive from outside
Seven coding-agent harnesses, such as Claude Code and Codex, come preinstalled. Your app can prompt them, steer them mid-task and stream their events through one API. With the optional, paid AI credentials proxy on, the real keys of supported model providers stay outside the sandbox.
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
Modelence
Modelence is a managed backend platform for TypeScript web apps, with an open-source framework and an App Builder that makes apps from a prompt.
Modelence Cloud builds a Docker image from your repository on an official Node.js base and runs your app's web process, with scheduled jobs through the Modelence framework and a MongoDB Atlas database wired in. It runs on AWS and restarts crashed processes, reroutes traffic and adds capacity under load. It is built for startups that want production hosting without a cloud engineer. Its docs use "sandbox" for a per-app test environment, not for a machine an agent works in.
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. “5th cheapest of 365” means 4 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 | boat: price rank | Modelence: price rank |
|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A software factory needs a machine where agents install tools, run Docker and keep their work. Modelence runs your Node.js app and its scheduled jobs, and its docs list Docker Compose and non-Node runtimes as not compatible. | 5th cheapest of 365 $634 a month | 280th of 365, with limits $4,224 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB Modelence keeps a web app running all month. Its docs describe no SSH, no container disk and no desktop that an assistant could use as its own workspace. | 6th cheapest of 361 $657 a month | 290th of 361, with limits $4,380 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Modelence documents no API to start a fresh container per job. To run code per job, you run it inside your own Node.js app. | 189th of 336, with limits $2,000 a month | 234th of 336, with limits $200 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Giving each of your users a sandbox means one isolated machine per user. Modelence scales the replicas of one app instead. | 3rd cheapest of 349 $209 a month | 256th of 349, with limits $1,395 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 39th of 349, with limits $2,000 a month | 263rd of 349, with limits $6,000 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month | 10th cheapest of 359 $200 a month | 352nd of 359, with limits $1,200 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI jobs that build containers or run test suites in other languages hit Modelence's rule: projects that need Docker Compose, a custom Dockerfile or a non-Node runtime are listed as not compatible. | 2nd cheapest of 374 $86 a month | 301st of 374, with limits $672 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box needs SSH and a disk that keeps your files. Modelence documents neither for its containers; its lasting data goes into the managed database and file storage. | 19th cheapest of 371 $26 a month | 291st of 371, with limits $175 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Computer-use agents need a desktop and a browser to drive. A desktop is outside Modelence's scope, and its docs list a custom Dockerfile, which you would need to add one, as not compatible. | 7th cheapest of 367 $200 a month | 284th of 367, with limits $1,200 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month | 7th cheapest of 356 $200 a month | 298th of 356, with limits $600 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month | 6th cheapest of 343 $90 a month | 293rd of 343, with limits $300 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month | 73rd of 327, with limits $1,800 a month | 173rd of 327, with limits $6,000 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 3rd cheapest of 351 $200 a month | 263rd of 351, with limits $300 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month | Not ranked: misses a requirement | 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 |
Key features
| Feature | boat | Modelence |
|---|---|---|
| Kind of product | Sandboxes: a full Linux VM per agent or per user | Production hosting for Node.js web apps, databases and scheduled jobs |
| Isolation | Full VM with its own kernel (x86_64, KVM) | Docker container per app on AWS |
| What you can run | Any Linux software on Ubuntu 24.04: any language, many processes, root and systemd; no GPU, and you cannot boot your own image | One Node.js web process per app (several replicas allowed), with scheduled jobs through the Modelence framework; imported apps that need a separate worker process, a non-Node runtime or Docker Compose are listed as not compatible |
| Docker and nested virtualization | Docker with Compose, and /dev/kvm, on every sandbox | Projects that need Docker Compose or a custom Dockerfile are listed as not compatible |
| Shell access | SSH and a command API, as root | No SSH documented; logs, metrics and settings in the dashboard |
| Disk | Disk saved every minute and at each stop, kept at least 14 days after a stop; writes after the last snapshot can be lost if a machine fails | Container rebuilt from the repository on each deploy; state lives in the managed database and file storage |
| Fork | Fork a running sandbox from its last 1-minute snapshot | No fork documented; you change the number and size of containers |
| Desktop and browser | 1920x1080, 60 fps desktop with Chrome and computer-use tools | Outside Modelence's scope; a custom Dockerfile to add one is listed as not compatible |
| Coding agents | Seven harnesses preinstalled, such as Claude Code and Codex, with a prompt, steer and events API | A Claude Code plugin, a docs MCP server and a Cloud MCP server let your own coding agents write, deploy and inspect Modelence apps; it does not run agents for you |
| Keeping keys out of the machine | Optional AI credentials proxy ($1 per day on, plus $1 per 1,000 requests): with it on, real keys of supported model providers stay out of the sandbox; other secrets still enter it | Secrets are environment variables injected into the container at start |
| Managed databases | None; run your own inside a sandbox | MongoDB Atlas database provisioned per environment, with PostgreSQL and Redis available |
| Running in production | You restart or resize sandboxes yourself through the API | Autopilot on Pro and above replaces crashed instances, reroutes traffic and adds capacity under load |
| Inbound access | Up to 50 public HTTPS URLs per sandbox, token-gated by default; public IPv6; no dedicated IPv4 | One public URL per environment, with unlimited custom domains on paid plans |
| Regions | Germany, Finland and France; you cannot pick a region, and snapshot copies can be stored outside the EU | Region chosen per production environment on AWS: 3 self-serve (Oregon, Canada, Frankfurt) and 14 more on request on Pro and Scale; test environments run in Oregon |
| 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 | Modelence |
|---|---|---|
| Full VM (own kernel) | Yes | No documented way |
| Docker inside | Yes | No documented way |
| Nested virtualization | Yes | No documented way |
| Root | Yes | Not published |
| Systemd | Yes | Not published |
| Desktop GUI | Yes | No documented way |
| Browser | Yes | No documented way |
| Computer-use API | Yes | No documented way |
| SSH | Yes | No documented way |
| Persistent disk | Yes | No documented way |
| Pause/resume | Yes | Not published |
| Automatic snapshots | Yes | No documented way |
| Fork/clone | Yes | No documented way |
| Live fork (no pause) | Yes | No documented way |
| Memory fork | No documented way | No documented way |
| Custom image (Docker or snapshot) | Yes | No documented way |
| ≥24 h sessions | Yes | Yes |
| Idle auto-stop | No documented way | No documented way |
| Wake on request | No documented way | No documented way |
| HTTPS preview URLs | Yes | Yes |
| Raw TCP inbound | Yes | Not published |
| Public IPv4 | No documented way | No documented way |
| Egress allowlist | No documented way | No documented way |
| Secret proxy | Yes | No documented way |
| Secret proxy for any API | Yes | No documented way |
| Preinstalled agents | Yes | No documented way |
| Agent harness API | Yes | No documented way |
| MCP server | No documented way | Yes |
| GPU | No documented way | No documented way |
| Windows | No documented way | No documented way |
| Arm64 | No documented way | No documented way |
| Self-hosting | No documented way | No documented way |
| BYOC | No documented way | No documented way |
| Open source | No documented way | Yes |
| Scoped API keys | Yes | Not published |
| Spend limits | Yes | Not published |
| Usage per sandbox (API) | Yes | No documented way |
| Audit logs | No documented way | Not published |
| SSO | No documented way | Yes |
| SOC 2 | No documented way | Yes |
| HIPAA | No documented way | No documented way |
| EU data residency | No documented way | Yes |
| Zero data retention | Yes | Not published |
Performance
Modelence is not in the public benchmarks we use, so we cannot say which is faster. On the V8 Web Tooling Benchmark at 4 vCPU / 8 GB, boat did 28.06 runs/s, second of 15, in StarSling's run of 3 October 2026 (run 37073680030). 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.
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. Modelence is 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 | boat, a month | Modelence, 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 | $42 · Starter · with limits |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $26 · $20/mo | $175 · Starter · 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 | $2,112 · Free · 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 | $600 · Free · with limits |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $8,800 · $2000/mo · with limits | $1,200 · Free · with limits |
boat: Plans cost $20, $100, $500 or $2,000 a month. The whole plan price comes back as VM time for that month, and it does not carry over. A default sandbox (4 vCPU / 8 GB) costs $0.036 an hour, billed per second while it runs. A stopped sandbox has no compute charge, and boat keeps its disk for at least 14 days; saved snapshots above the included count and the AI credentials proxy still charge while sandboxes are stopped. Each plan includes a number of starts (the $20 plan: 200 a day, 60 an hour, 12 a minute). Extra starts cost $0.03 to $0.07 each, and starts above 6 times the included number (3 times per minute) fail.
Modelence: Modelence bills containers by the hour based on their size, from Micro (0.25 vCPU) to X-Large Plus (4 vCPU / 16 GB). The 4 vCPU / 8 GB size used here is $0.24 an hour. Paid plans (Starter $20, Pro $100, Scale $500 a month) include the same amount as cloud usage, and extra usage is billed at the end of the month. Enterprise is a yearly contract from $12,000. The pricing page does not say whether a stopped environment bills. Disk is not priced separately, so our bills leave it out.
Limits in those scenarios
Every scenario is marked not a clean fit for Modelence. It hosts Node.js apps per environment rather than handing out machines, its containers have no documented disk of their own, and Docker Compose is listed as not compatible. Its plans also cap total vCPU across your apps: 4 on Starter, 16 on Pro and 64 on Scale, so 50 agents at 4 vCPU each need a custom Enterprise limit. On boat, the 1,000-job burst is the one that bites: the $2,000 plan includes 90 starts a minute, 600 an hour and 2,400 a day, and the job needs about 1,000 an hour and 6,667 a day. Starts above the included numbers cost $0.03 to $0.07 each, and starts above 270 a minute fail. boat also has no region choice, and a machine failure loses the writes made after the last 1-minute snapshot.
| Scenario | boat | Modelence |
|---|---|---|
| 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 |
|
| 1,000 short jobs at once |
|
|
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
Which one should you use?
Use Modelence when
you want a Node.js web app, its database and scheduled jobs hosted and kept running for you, especially in a US, Canadian or other AWS region.
- Hosting your web app in production. Useful for a Node.js backend that must stay up: crashed instances are replaced and capacity is added under load.
- Managed databases included. Useful if you want a database without running one: each environment gets a MongoDB Atlas database, with PostgreSQL and Redis available.
- A choice of regions. Useful if your users are in the US or elsewhere: you pick Oregon, Canada or Frankfurt, and 14 more AWS regions on request on Pro and Scale.
Use boat when
an agent needs its own machine to work in: any language, Docker, root, a disk saved every minute, a desktop, and coding agents you drive from outside, with an optional proxy that keeps model keys out.
- A whole machine, not one web app. An agent that runs pip install, docker compose or a Go test suite needs more than a Node.js app. Modelence's deploy docs list a separate worker process, non-Node runtimes and Docker Compose as not compatible.
- A disk that keeps the agent's work. Agents build up state: cloned repos, installed packages, half-done edits. On Modelence, that state lasts only if your code writes it to the database or file storage.
- Coding agents you drive from outside. Modelence's Claude Code plugin, docs MCP server and Cloud MCP server let your own agents write, deploy and inspect Modelence apps. To run agents for your users there, you build the harness yourself, and keys are environment variables the agent's code can read.
Questions people ask
Is Modelence a sandbox for AI agents?
No. Modelence is a platform for hosting Node.js web apps with managed databases, plus an App Builder that makes apps from a prompt. Its docs use "sandbox" for a per-app test environment. It runs your app's web process in a container, and it does not hand out machines for agents to work in. boat does that.
Is boat cheaper than Modelence?
For four of the five scenarios on this page, at list prices, yes. One agent on workdays is $20 on boat and $42 on Modelence. One agent running all month is $26 against $175. 50 agents are $317 against $2,112, and Modelence needs a custom Enterprise limit for them. Sandboxes for your users are $200 against $600. For 1,000 short jobs at once, Modelence prices lower, $1,200 against $8,800, but neither fits as listed: Modelence's self-serve plans cap total vCPU at 64, and boat refuses starts above 270 a minute.
Can I run Docker or Python on Modelence?
Not on Modelence Cloud as documented. Its deploy docs say the image runs Node.js only, and list projects that need Docker Compose, a custom Dockerfile or a non-Node runtime as not compatible. A boat sandbox runs any Linux software, with Docker and root.
Does Modelence keep files between runs?
Its containers are rebuilt from your repository on each deploy, and lasting data goes into the managed MongoDB database or Modelence's file storage. A boat sandbox saves its disk every minute and at each stop, and keeps it for at least 14 days after a stop. Memory, running programs and /tmp are not saved, and writes after the last snapshot can be lost if a machine fails.
Where does my data run?
Modelence runs on AWS in the region you choose per production environment, the United States by default. Frankfurt is one of the 3 self-serve regions, and test environments run in Oregon. boat sandboxes run on machines in Germany, Finland and France, and you cannot pick a region. Snapshot copies can be stored outside the EU.
Is boat SOC 2 compliant?
SOC 2 is in progress, and boat has no SOC 2 report today. Modelence lists SOC 2 on its Enterprise plan. Live status is at status.ascii.dev and our data retention policy is at docs.boat.dev/data-retention.
Start a sandbox
Install the CLI, then run boat new.
7-day free trial