llms.txt
+200 −51
https://mobilebox.dev/llms.txt
− > Cloud Android phones for AI agents. Mobilebox runs Android 14 arm64 virtual sandboxes that your agent controls over a REST API or a hosted MCP server with 12 tools. The agent reads the screen as numbered elements, acts with taps, typing and swipes, installs APKs, and parks the phone to keep apps, data and logins. Your agent supplies the reasoning; Mobilebox supplies the phone.− ## Pricing and limits− - $0.01 per ready phone-minute, billed per second, one-minute minimum per session. Boot time and parked phones are free.− - Sign up with GitHub at https://mobilebox.dev/signup; new accounts start with $2 of credit (200 phone-minutes).− - Up to 3 running and 10 kept phones per workspace, shared by its keys. Up to 20 actions per call, 256 KB of JSON, waits up to 55 s.− - Phones park after 5 idle minutes, and after 15 minutes of running unless max_duration asks for up to 3 hours.− - Not available: Google Play, SMS, public ADB or shell.− - Early access: iPhone/iPad and real Android devices. Create with platform: ios is answered 409 early_access (no phone, no charge); device: real gets a cloud phone with a notice, billed as cloud. Both put the account on the list: https://mobilebox.dev/docs/reference#early-access-devices.− - MCP: https://mobilebox.dev/mcp (stateless Streamable HTTP; header Authorization: Bearer API_KEY). Tools: list_phones, create_phone, start_phone, park_phone, delete_phone, set_device, install_app, list_apps, uninstall_app, observe, act, screenshot.− - REST: https://mobilebox.dev/v1 with the same key; schemas in [OpenAPI](https://mobilebox.dev/openapi.json).− - Agent keys come from the dashboard after sign-up. Public discovery does not authenticate the client. Never put keys in prompts.− - [How can Claude Code test my Android app?](https://mobilebox.dev/blog/claude-code-android-app-testing)− - [How do I run an Android emulator in CI without KVM?](https://mobilebox.dev/blog/android-emulator-without-kvm)− - [How do I run an APK in the cloud via an API?](https://mobilebox.dev/blog/android-emulator-without-kvm)− - [Which cloud Android phones can AI agents use (Mobilerun or Limrun alternatives)?](https://mobilebox.dev/blog/cloud-android-for-ai-agents-compared)− - [Which Android MCP server should I use, hosted or local (mobile-mcp)?](https://mobilebox.dev/blog/android-mcp-server-hosted-vs-local)− - [How do I give my coding agent an Android phone over MCP?](https://mobilebox.dev/blog/android-agent-mcp)− - [How do I smoke-test an APK in GitHub Actions without a local emulator?](https://mobilebox.dev/blog/android-apk-ci-smoke-tests)− - [Quickstart](https://mobilebox.dev/docs/index.md): Runnable Node lifecycle and authentication.− - [Agent guide](https://mobilebox.dev/skill.md): Observe, act, verify, park.− - [API reference](https://mobilebox.dev/docs/reference.md): REST endpoints, exact limits and response semantics.− - [Errors](https://mobilebox.dev/docs/errors.md): Recovery and partial failures.− - [OpenAPI](https://mobilebox.dev/openapi.json): REST schemas including the shared action vocabulary.− - [Give your coding agent an Android sandbox over MCP](https://mobilebox.dev/blog/android-agent-mcp): Add the remote MCP server to Claude Code or Cursor, see every phone tool, run a bounded first task and know what is proven.− - [Does this task need a phone or a browser?](https://mobilebox.dev/blog/native-app-or-browser): Choose the smallest runtime that can actually complete the task: HTTP, browser, or native Android.− - [Mobilebox or Mobilerun: where do the agent and the phone run?](https://mobilebox.dev/blog/mobilebox-vs-mobilerun): Compare Mobilebox’s hosted Android API and MCP with Mobilerun’s framework, coding-agent harness, hosted agent and cloud phones.− - [Mobilebox or Appium for Android automation?](https://mobilebox.dev/blog/mobilebox-vs-appium): Separate a managed phone runtime from a test automation framework before choosing your Android stack.− - [Android, E2B, Browserbase: which sandbox fits?](https://mobilebox.dev/blog/android-vs-agent-sandboxes): Compare execution surfaces and state, rather than treating every agent sandbox as interchangeable.− - [Why your agent tapped the wrong thing](https://mobilebox.dev/blog/screen-refs-and-recovery): Understand stale screen references, choose stable targets, and recover without repeating side effects.− - [Batch the steps you know. Observe the ones you don’t.](https://mobilebox.dev/blog/batch-android-actions): Build bounded Android action batches, inspect partial failure, and verify the resulting screen.− - [Install APKs and split bundles over an API](https://mobilebox.dev/blog/install-apk-split-bundles): Upload APK bytes or install by URL, handle arm64 split bundles, and diagnose update failures.− - [Keep the login. Stop the running phone.](https://mobilebox.dev/blog/park-resume-android): Understand what park and resume preserve, what restarts, and when to delete an Android sandbox.− - [Put a budget around your Android agent](https://mobilebox.dev/blog/android-sandbox-costs): Use idle timeouts, session caps, account balance, and session records to control phone runtime costs.− - [Timezone, locale, GPS and IP are four different inputs](https://mobilebox.dev/blog/locale-location-proxy): Configure an Android sandbox for regional workflows without confusing locale or GPS with network egress.− - [A successful tap is not a successful task](https://mobilebox.dev/blog/verify-agent-workflows): Design Android agent checks around outcomes, screen evidence, failure traces, and safe cleanup.− - [Screenshot or accessibility tree? Choose the evidence your agent needs](https://mobilebox.dev/blog/screenshot-vs-accessibility-tree): Choose structured Android observations for labels and state, screenshots for pixels, and synchronized evidence when both matter.− - [Run an APK smoke test from CI without a local emulator](https://mobilebox.dev/blog/android-apk-ci-smoke-tests): A complete Node script and GitHub Actions job that installs an APK on a remote Android sandbox, checks its first screen and deletes the phone even when the check fails.− - [Mobilebox or BrowserStack: agent runtime or device lab?](https://mobilebox.dev/blog/mobilebox-vs-browserstack): Compare an agent-facing Android sandbox with a real-device Appium testing workflow, using documented responsibilities instead of rankings.− - [Mobilebox or Genymotion: compare the Android control surface](https://mobilebox.dev/blog/mobilebox-vs-genymotion): Compare device provisioning, ADB-based toolchains, agent actions and the exact persistence semantics of your chosen Android workflow.− - [Give each agent a key. Keep the workspace boundary clear.](https://mobilebox.dev/blog/android-api-key-security): Understand project-wide API-key access, browser sessions, safe rotation and the difference between a named key and an isolated project.+ > Mobile sandboxes for AI agents. Mobilebox runs cloud Android 14 phones (arm64), iPhone and iPad, and real Android devices that your agent controls over a REST API or a hosted MCP server with 15 tools. The agent reads the screen as numbered elements, acts with taps, typing and swipes, installs APKs, and parks the phone to keep apps, data and logins. Your agent supplies the reasoning; Mobilebox supplies the phone.+ ## Pricing+ - Cloud Android phone: $0.01 per ready phone-minute, billed per second, one-minute minimum per session. Boot time and parked phones are free.+ - Real Android device: $0.01 per phone-minute, the cloud price. iPhone and iPad: $0.05 per phone-minute. Dedicated devices from $79 a month; volume over 100 devices through sales.+ - Your agent sets you up: [setup.md](https://mobilebox.dev/setup.md?via=llms) (your email, and the agent gets its own key at once; the 6-digit code from your inbox keeps the workspace). Or sign up at https://mobilebox.dev/signup?via=llms. New workspaces get free credit. Add credits from the dashboard.+ - Requests: up to 20 actions per call, 256 KB of JSON, waits up to 55 s. Phones park after 5 idle minutes, and after 15 minutes of running unless max_duration asks for up to 3 hours.+ - iPhone and iPad: create with platform: "ios". Real Android: device: "real". Pages: [iPhone and iPad](https://mobilebox.dev/ios?via=llms), [real Android devices](https://mobilebox.dev/real-devices?via=llms). Details: https://mobilebox.dev/docs/reference?via=llms#devices.+ - MCP: https://mobilebox.dev/mcp (stateless Streamable HTTP; header Authorization: Bearer API_KEY). Tools: list_phones, create_phone, start_phone, park_phone, delete_phone, set_device, install_app, list_apps, uninstall_app, observe, act, screenshot, watch_phone, publish_proof, request_capability.+ - One-click: add https://mobilebox.dev/mcp as a connector in Claude, ChatGPT, Cursor or Claude Code with no key; the client signs your user in over MCP OAuth ([how](https://mobilebox.dev/docs/reference?via=llms#mcp-sign-in)).+ - Setup per coding tool: [Claude Code](https://mobilebox.dev/blog/android-agent-mcp?via=llms), [Cursor, Codex and VS Code](https://mobilebox.dev/blog/android-mcp-cursor-codex-vscode?via=llms), [Gemini CLI](https://mobilebox.dev/blog/gemini-cli-android-mcp?via=llms), [Windsurf](https://mobilebox.dev/blog/windsurf-android-mcp?via=llms), [Cline](https://mobilebox.dev/blog/cline-android-mcp?via=llms), [GitHub Copilot](https://mobilebox.dev/blog/github-copilot-android-mcp?via=llms).+ - REST: https://mobilebox.dev/v1 with the same key; schemas in [OpenAPI](https://mobilebox.dev/openapi.json?via=llms).+ - Python: `pip install mobilebox` is the REST client ([PyPI](https://pypi.org/project/mobilebox/)), `pip install langchain-mobilebox` its LangChain and LangGraph tools ([PyPI](https://pypi.org/project/langchain-mobilebox/)); both read MOBILESANDBOX_API_KEY. How-to: [an Android phone for your Python AI agent](https://mobilebox.dev/blog/android-phone-for-python-ai-agent?via=llms).+ - Agent keys: the agent gets its own through [setup.md](https://mobilebox.dev/setup.md?via=llms) (POST /v1/signup/email returns it at once; /v1/signup/verify with the emailed code confirms it), or your user creates one in the dashboard. Public discovery does not authenticate the client. Never put keys in prompts.+ ## Pages+ - [Mobile sandbox for AI agents](https://mobilebox.dev/mobile-sandbox?via=llms): what a mobile sandbox is and how an agent uses one.+ - [Mobile Sandbox für KI-Agenten](https://mobilebox.dev/de/mobile-sandbox?via=llms): the mobile sandbox page in German.+ - [AI sandboxes for agents](https://mobilebox.dev/ai-sandbox?via=llms): code, browser or phone sandbox, and which one a task needs.+ - [Phone agents](https://mobilebox.dev/phone-agent?via=llms): give an AI agent a phone it can use (not a voice-call agent).+ - [App automation with AI agents](https://mobilebox.dev/app-automation?via=llms): test your own app, or run tasks in apps with your own account.+ - [Android MCP server](https://mobilebox.dev/android-mcp-server?via=llms): the hosted Android MCP server, how to connect any client, every tool, hosted or local.+ - [Live view](https://mobilebox.dev/live-view?via=llms): watch your AI agent use a phone live, every tap, swipe and keystroke drawn and captioned. Open it with watch_phone, POST /v1/phones/{id}/view or npx mobilebox watch <phone_id>.+ - [iPhone and iPad](https://mobilebox.dev/ios?via=llms) and [real Android devices](https://mobilebox.dev/real-devices?via=llms): the same API and MCP on every device.+ - [Changelog](https://mobilebox.dev/changelog?via=llms): what shipped, newest first, dated.+ - [About](https://mobilebox.dev/about?via=llms): who builds Mobilebox, contact and support.+ - [Benchmarks](https://mobilebox.dev/benchmarks?via=llms): measured cold start, tap latency, APK install and 10-minute cost for cloud Android services and a CI emulator (Nov 2026), with CSV and scripts.+ ## Integrations+ Each page has a runnable example (the code creates a phone over REST, the agent drives it, the phone is deleted in finally). All: [Integrations](https://mobilebox.dev/integrations?via=llms).+ - [OpenAI Agents SDK](https://mobilebox.dev/integrations/openai-agents?via=llms): Python, mCP through MCPServerStreamableHttp. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/openai-agents/agent.py+ - [LangGraph](https://mobilebox.dev/integrations/langgraph?via=llms): Python, the langchain-mobilebox tools over REST. Example: https://github.com/fifteenA/mobilebox/tree/main/packages/python/langchain-mobilebox+ - [CrewAI](https://mobilebox.dev/integrations/crewai?via=llms): Python, mCP through MCPServerHTTP. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/crewai/agent.py+ - [browser-use](https://mobilebox.dev/integrations/browser-use?via=llms): Python, two custom actions over REST. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/browser-use/agent.py+ - [Mastra](https://mobilebox.dev/integrations/mastra?via=llms): TypeScript, mCP through MCPClient. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/mastra/agent.ts+ - [Vercel AI SDK](https://mobilebox.dev/integrations/vercel-ai-sdk?via=llms): TypeScript, mCP through createMCPClient. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/vercel-ai-sdk/agent.ts+ - [Claude Agent SDK](https://mobilebox.dev/integrations/claude-agent-sdk?via=llms): TypeScript, mCP through mcpServers. Example: https://github.com/fifteenA/mobilebox/blob/main/examples/integrations/claude-agent-sdk/agent.ts+ - [How can Claude Code test my Android app?](https://mobilebox.dev/blog/claude-code-android-app-testing?via=llms)+ - [How do I run an Android emulator in CI without KVM?](https://mobilebox.dev/blog/android-emulator-without-kvm?via=llms)+ - [How do I run an APK in the cloud via an API?](https://mobilebox.dev/blog/android-emulator-without-kvm?via=llms)+ - [Which cloud Android phones can AI agents use (Mobilerun or Limrun alternatives)?](https://mobilebox.dev/blog/cloud-android-for-ai-agents-compared?via=llms)+ - [Which Android MCP server should I use, hosted or local (mobile-mcp)?](https://mobilebox.dev/blog/android-mcp-server-hosted-vs-local?via=llms)+ - [How do I give my coding agent an Android phone over MCP?](https://mobilebox.dev/blog/android-agent-mcp?via=llms)
Showing the first 40 of each.