VibeView and Limrun both give a coding agent a cloud iOS simulator and Android emulator, driven through a CLI, a skill and an MCP server, with builds on a hosted Mac and uploads to TestFlight and Google Play. VibeView covers more of the job around that loop: Apple TV and Android TV on cloud devices and Roku on real hardware, flows that replay as tests with visual regression and a CI check, and a dashboard where someone who never opens a terminal can record one.
And it costs less to use every day. VibeView includes device minutes in each plan at the same rate on iOS, Android and TV; Limrun bills credits, and an iOS minute costs twice an Android one. At 20,000 iOS minutes a month, Limrun costs $1,950 and VibeView $249, so Limrun costs about 7 times as much.
Two commands, then ask your agent to verify something
npm install -g vibeview, then vibeview agent-setup. The skill
lands where Claude Code finds it and the MCP server is one prompt away. Limrun's equivalent
is lim skills install; if you have run one, you can run the other.
What a month of device time costs
The cheapest way to buy each amount on each side: a plan, plus overage where the plan runs out. VibeView minutes count the same on iOS, Android and TV. Monthly prices; Limrun's annual plans save 20% and ours about 16%.
| Device minutes a month | VibeView (any platform) | Limrun, iOS simulator | Limrun, Android emulator |
|---|---|---|---|
| 500 | $19 — Solo | $45 — Entry + overage | $20 — Entry |
| 3,000 | $49 — Team | $250 — Individual + overage | $145 — Entry + overage |
| 20,000 | $249 — Business | $1,950 — Individual + overage | $950 — Individual + overage |
Builds come on top: every plan includes build minutes (150 on Solo, 2,000 on Business) besides its device minutes, and past them a Mac build minute is $0.08 and a Linux one $0.03. On Limrun, builds draw on the same credits as devices: an active Xcode minute is $0.08 inside a plan and $0.10 past it.
Limrun alternatives: how VibeView compares
| What you are deciding | VibeView | Limrun |
|---|---|---|
| Does it cover TV apps? | Yes — tvOS and Android TV run on cloud TV devices with a d-pad and a focus walk the agent can call; Roku runs on real hardware | tvOS builds, but "tvOS and visionOS are device builds only", so there is no tvOS simulator to run them on; Android TV and Roku are not in their documentation |
| Does anything test the app after the agent leaves? | Yes — record a flow in the browser and replay it on any platform of the same form factor, compare screenshots after every step (on by default in every new group), hand a moved button to an AI agent (beta), and run it from CI with a GitHub check | Runs Maestro, Appium, XCTest and other frameworks you bring; a recorder, replay or visual comparison is not in their documentation |
| What does a month of daily use cost? | $249 for 20,000 device minutes on any platform (Business). Free: 30 minutes and 30 build minutes a month, no card | $1,950 for 20,000 iOS minutes, $950 for Android. Trial: 120 one-time credits |
| Can non-engineers use it? | Yes — share links that need no account, live collaboration where a guest takes the controls, an embeddable player, a run link on pull request builds, and a dashboard where a tester records a test by walking through the app | Preview links on pull requests and a <RemoteControl /> component for embedding the simulator |
| Which phone devices does it run? | iOS simulators (iPhone, iPad, iPhone Duo) and Android emulators (phone, tablet, foldable), streamed to the browser | Cloud iOS simulators (iPhone, iPhone Duo, iPad and Watch models) and Android emulators (phone, tablet), streamed to the browser |
| How does the agent read the screen? | A ref-annotated UI tree — every label, id and rect as text with stable @e references — plus screenshots on demand | lim ios element-tree, picking elements by accessibility ID, plus screenshots |
| How does the agent act? | tap by ref or coordinates, type, scroll-to with end-of-list detection, drag, press, iOS alert handling, open-url; every action returns a diff of what changed | Accessibility-tree actions or "your favorite framework" (Appium, Maestro, Detox, agent-device); batch actions with millisecond precision |
| How does the agent get set up? | vibeview agent-setup installs the skill and offers to register the MCP server; also npx skills add vibeview/skills | lim skills install; a per-instance MCP server, a remote MCP server, and SDKs in TypeScript, Python and Go |
| Can it build the app without a Mac? | Yes — cloud builds on a managed Mac or Linux builder from one CLI command, with generated Apple and Android signing credentials | Yes — a hosted Xcode sandbox and Android builds, with Apple cloud signing, Bazel remote execution and remote cache |
| Can it submit to the stores? | Yes — vibeview submit to TestFlight and your tester groups for iOS and tvOS, and to a Google Play track or the Amazon Appstore (Fire TV) for Android and Android TV | Yes — App Store Connect and TestFlight for iOS and tvOS, Google Play's internal track for Android |
Three rows decide it: TV, what tests the app after the agent leaves, and what a month of daily use costs. What we don't do yet: on-demand video of a live session, camera simulation, per-instance agent credentials and region pinning — and our plans run 2 to 10 devices at once, where Limrun sets no limit beyond abuse-prevention quotas.
Where the loop ends
Both products get an agent onto a device, build the app and send it to the stores. The question this page turns on is what happens around that loop.
Before the device: both build on a hosted Mac and sign in the cloud. VibeView also generates and stores the Apple certificate, provisioning profile and Android keystore for your organization, so the credential a teammate needs is in the account rather than on your laptop.
After the agent: on VibeView, the flow the agent — or a tester — just walked can be saved as a test that replays on any platform of the same form factor, compares screenshots after every step (on by default in every new group), and runs from CI with a GitHub check. When a button moves, the step is handed to an AI agent (beta) that finds it by name. Limrun runs the testing framework you bring — Maestro, Appium, XCTest — and leaves writing and maintaining those tests to you.
And on TV: the agent verbs are the same ones, with a focus walk that moves the d-pad to a named element. Limrun compiles tvOS apps, but by its own documentation only as device builds, with no tvOS simulator to run them on.
Which fits you: a decision guide
You ship a TV app
VibeView. tvOS and Android TV on cloud devices, Roku on real hardware. Limrun compiles tvOS but has no tvOS simulator.
You want tests that outlive the agent
VibeView. Recorded flows, visual regression, a CI runner and a GitHub check, in the product.
Your devices run all day
VibeView. 20,000 device minutes a month cost $249 on any platform; on Limrun the same iOS minutes cost $1,950.
Your testers don't use a terminal
VibeView. Share links with no account, live collaboration where a guest takes the controls, and a dashboard where a tester records a test by walking through the app.
You test on every platform
VibeView. One minute costs the same on iOS, Android and TV, so a cross-platform run doesn't cost double on iOS.
Your agent works through pull requests
Both. Limrun puts simulator previews on pull requests; VibeView's GitHub App builds each pull request and comments a link to run each build on a cloud device.
Run the same task on both and compare the transcript
Ask your agent to verify one change on a VibeView device and one on Limrun's. Read what it saw, what it did, and what it could prove. The transcript is the comparison; this page is only the map.
Frequently asked questions
How is VibeView different from Limrun?
They overlap more than any other pair on this site: both run cloud iOS simulators and Android emulators a coding agent can drive through a CLI, a skill and an MCP server, both build apps on a hosted Mac, and both upload to TestFlight and Google Play. VibeView additionally runs Apple TV and Android TV, and Roku on real hardware; records flows as replayable tests with visual regression and a CI runner; and costs less per device minute once you use it every day. Limrun adds camera simulation, per-instance agent credentials and region pinning, and sets no limit on devices at once beyond abuse-prevention quotas.
Can Limrun run tvOS apps?
Their Xcode build page says "tvOS and visionOS are device builds only", so a tvOS app can be compiled but not run on a Limrun simulator. Android TV and Roku are not in their documentation. VibeView runs tvOS and Android TV on cloud TV devices with a d-pad and a focus walk the agent can call, and Roku on real hardware.
Do both submit to the App Store and Google Play?
Yes. Limrun uploads iOS and tvOS builds to App Store Connect and TestFlight and publishes Android builds to Google Play's internal track. VibeView's vibeview submit sends signed builds to TestFlight for iOS and tvOS and on to your tester groups, with Beta App Review for external groups, and to a Google Play track or the Amazon Appstore for Android and Android TV. In both, App Store review submission stays in App Store Connect.
What does Limrun cost compared with VibeView?
Limrun bills credits: $0.04 each inside a plan and $0.05 past it, with an iOS simulator minute at 2 credits and an Android emulator minute at 1. Entry is $20 a month for 500 credits and Individual $200 for 5,000. VibeView bills device minutes the same on every platform: Solo is $19 a month with 1,000 minutes, Business $249 with 20,000, and past that $0.02 a minute. For 20,000 iOS minutes a month that is $249 on VibeView and $1,950 on Limrun. Limrun has no concurrency limit beyond abuse-prevention quotas; VibeView plans run 2 to 10 devices at once.
Which one has better agent controls?
They are close, and the details decide it. Limrun's agent picks elements from an accessibility tree, batches actions for millisecond timing, records 60fps video and can be sandboxed with a per-instance URL and token. VibeView's agent works from a ref-annotated UI tree where every action returns a diff of what changed, has scroll-to with end-of-list detection, handles iOS system alerts, reads app logs, and on TV can move focus to a named element. If you need region pinning or per-instance credentials, Limrun documents both and we do not.
Can non-engineers use either one?
Both share a session by URL: Limrun has preview links on pull requests and a RemoteControl component for embedding; VibeView has share links that need no account, live collaboration where a guest can take the controls, an embeddable player, run links on pull request builds, and a dashboard where a tester records a test by walking through the app. VibeView is the one built for the person who will never open a terminal.
Published 20 September 2026 · Updated 25 October 2026 · Limrun pricing page and documentation verified 23 October 2026