Is there an official Claude mobile app for coding?

Anthropic's own Claude app does chat well but doesn't host code, deploy apps, or persist agent state across sessions. VibeKit fills that gap: a native iOS app that uses Claude (your key or your Pro/Max subscription) as the underlying model, plus hosting + GitHub + persistent context per app.

Can I use my Claude Pro / Max subscription on mobile?

Yes — VibeKit supports Claude OAuth, the same auth path Claude Code uses. Sign in with Anthropic in Settings → AI Provider, and the agent runs against your Pro or Max subscription quota at no extra cost. No API key needed, no second AI subscription.

What can the Claude agent actually do from mobile?

Same as Claude Code on a Mac: edit files, run builds, debug runtime errors, commit to GitHub, deploy your app. The differences are (a) the workspace lives in VibeKit's cloud, not your laptop, so it persists across devices, and (b) you direct the agent via chat from iOS instead of typing in a terminal.

Is it cheaper than using Claude Code on a laptop?

Same model price either way — Anthropic bills you per token regardless of which client makes the call. With a Pro or Max subscription, BYOK on either platform is effectively flat-rate. VibeKit adds hosting (1 app free), GitHub integration, and the iOS app — Claude Code adds nothing beyond the agent loop on your own machine. They're complementary.

Claude on Mobile

VibeKit is the most complete way to use Claude for real coding work on iPhone. Claude Sonnet powers the agent — via your sk-ant- API key or, better, your Claude Pro / Max OAuth subscription. Builds, debugs, commits to GitHub, and deploys to a live URL. From your phone.

What you're actually getting

Anthropic's own Claude app is great for chat but doesn't touch your code, host anything, or deploy. Claude Code is a CLI — terminal-only, your laptop only. VibeKit is the layer in between: a native iOS app where Claude has tools, a real container, and persistent context per app.

Claude Sonnet as the agent. Same model. Same quality. Same tool-use loop you'd get on Claude Code.
OAuth or API key. Sign in with your Claude Pro / Max subscription (no per-token billing) or paste a sk-ant- key for usage-based.
Persistent across sessions. Your agent remembers what's in your workspace and what changed yesterday. No re-pasting context every time.
Real container, real URL. Each app lives at yourname.vibekit.bot with auto-SSL.

BYOK options for Claude

Option	Best for	Cost shape
Claude Pro / Max OAuth	Heavy users on a subscription	Flat $20 or $100/mo (Anthropic), $0 extra
`sk-ant-` API key	Light or bursty usage	Per-token, no minimum
Platform fallback (no BYOK)	Trying it out	OpenRouter Claude price + 20%

If you already pay for Claude Pro, OAuth is a strict upgrade — same allowance, same model, no extra $/token on top, just channeled through VibeKit's agent loop.

How it differs from Claude Code on a Mac

Functionally similar — same agent loop, same model, same kinds of tools. The differences are device + persistence:

Mac: Code lives on your laptop. Agent state is local. When you close the laptop, the agent goes with it.
VibeKit: Code lives in the cloud, attached to the app's container. Agent state persists across devices and sessions. Open the iOS app or the Telegram bot or the web dashboard — same agent, same context.

Most real users end up using both: Claude Code on the Mac for deep work, VibeKit on mobile for "fix this 500 I just noticed" while away from the laptop.

Setup

Install VibeKit from the App Store (iOS)
Sign in with Apple or Google
Settings → AI Provider → Sign in with Claude (or paste sk-ant-…)
Pick a template or connect a GitHub repo
Chat with Claude. Tap your URL. Done.

Claude on iPhone — same model, real deploys

Use your Claude Pro / Max subscription via OAuth. Zero markup.

Download on the App Store →