LibreAI / KI-Agenten / agent-device
// agents · callstack/agent-device

agent-device

agent-device ist ein Open-Source-Projekt aus der Kategorie KI-Agenten von callstack (callstack/agent-device auf GitHub). Es hat 4.936 Sterne auf GitHub und 328 Forks, wird unter der Lizenz MIT veröffentlicht und wurde zuletzt am 8. Oktober 2026 auf GitHub aktualisiert.

Mobile app automation and verification built for AI coding agents — CLI, MCP server, and Node API

★ 4,9 Tsd. Sterne⑂ 328 ForksMITaktualisiert Okt 2026
▦ Visual — das README dieses Projekts enthält keinen einbettbaren Screenshot,
darum hier ein Platzhalter. Die Anleitung unten ist trotzdem vollständig.
// projektdaten

Projektdaten

GitHub-Sterne
4.936
Forks
328
Lizenz
MIT
Zuletzt auf GitHub aktualisiert
8. Oktober 2026
Kategorie
KI-Agenten
Eintragstyp
Vollständige Installationsanleitung
Offizielle Website
Official site
// so funktioniert es

So funktioniert es

  1. Check your goal

    Decide which apps and platforms your agent must verify — iOS simulator, Android emulator, a physical device, web, or desktop.

  2. Install the CLI

    Install agent-device globally from npm; it requires Node.js 22.12 or newer (Node 24+ for web automation).

  3. Verify your setup

    Run the doctor command yourself before handing the CLI to an agent, and review the workflow guides.

  4. Run your first session

    Open an app, take an accessibility snapshot, act on a ref, and capture a screenshot as evidence.

// funktionen

Warum es so beliebt ist

Accessibility snapshots

Agents inspect app state through token-efficient accessibility trees, refs, and selectors rather than screenshots alone.

Act on the UI

Tap or press elements, fill fields, scroll, use gestures, wait, assert state, and handle alerts.

Debugging evidence

Capture screenshots, video, logs, traces, network data, performance samples, and crash details.

Replay scripts

Save working steps as .ad scripts for local use or CI, with export to strict Maestro YAML when needed.

MCP server built in

`agent-device mcp` starts a stdio MCP server exposing the same commands as structured tools.

Parallel-safe sessions

Sessions are scoped to the caller's git worktree, and host-local device claims stop parallel agents from taking over each other's simulators; remote device clouds such as BrowserStack and AWS Device Farm are supported.

Advertisement
// selbst ausführen

Führe es selbst aus

  1. Clone the repository

    Clone the project and enter its directory (or skip the clone and install from npm in the next step).

    git clone https://github.com/callstack/agent-device.git
    cd agent-device
  2. Install the CLI

    Install agent-device globally; Node.js 22.12+ is required.

    npm install -g agent-device@latest
  3. Check setup and drive an app

    Run the doctor check, then open an app and inspect it.

    agent-device doctor
    agent-device open Contacts --platform ios
    agent-device snapshot -i
// anwendungsfälle

Was du damit bauen kannst

Also on this shelf: AgentArk — a personal AI assistant that puts security first. The AI Agents shelf also holds sglang — High-performance serving for the LLMs that power your agents. Readers working through the AI Agents shelf also stop at chroma, open-source search infrastructure built for AI workloads.

Advertisement
// häufige fragen

Häufige Fragen

Who maintains agent-device?

Callstack, the React Native consultancy; the README notes teams at companies such as Expensify and Shopify use it.

Is agent-device free?

Yes, the tool itself is open source under the MIT license. Optional remote device clouds it can connect to (such as BrowserStack or AWS Device Farm) are separate paid services.

Does it need a paid API key?

No API key is required for the CLI, MCP server, or Node API themselves; you bring your own coding agent (for example Claude Code or Codex) to drive it.

How does it compare to Appium, Detox, or Maestro?

Those tools run test suites that humans write and maintain; with agent-device, an agent reads app state and chooses each command at run time. It complements them — runs can be saved as .ad scripts or exported as strict Maestro YAML.

How do I get started?

Run `npm install -g agent-device@latest`, then `agent-device doctor`, open an app with `agent-device open`, and inspect it with `agent-device snapshot -i`.

// weiter entdecken

Ähnliche Tools

KI-Agenten

agent-browser

A fast browser-automation CLI built for AI agents.

★ 43,6 Tsd. SterneApache-2.0
Anleitung ansehen →
KI-Agenten

oh-my-agent

Open ai agents for mechanical verification for ai coding agents — skills p.

★ 1,3 Tsd. SterneMIT
Anleitung ansehen →
KI-Agenten

matlab-mcp-server

Official MathWorks MCP server: drive MATLAB from AI coding agents.

★ 1,6 Tsd. SterneNOASSERTION
Anleitung ansehen →
KI-Agenten

deep-research

Deep Research with any LLM: in-depth reports in minutes, with an SSE API and MCP server built in

★ 4,7 Tsd. SterneMIT
Anleitung ansehen →
KI-Agenten

chrome-devtools-mcp

Chrome DevTools for coding agents: an MCP server that lets agents debug and inspect real Chrome

★ 53,1 Tsd. SterneApache-2.0
Anleitung ansehen →
KI-Agenten

brightdata-mcp

Bright Data's MCP server: web search, scraping, structured data and browser automation for agents (75 tools)

★ 2,7 Tsd. SterneMIT
Anleitung ansehen →
// alternativen

Alternativen im Bereich KI-Agenten

Die anderen Projekte der Kategorie KI-Agenten mit den meisten Sternen im LibreAI-Katalog — die realistische Auswahl, falls agent-device nicht das Richtige ist.

// mehr in agents

Mehr in KI-Agenten

Verwandte Projekte der Kategorie KI-Agenten im Katalog, die agent-device bei den Sternen am nächsten stehen — zum Weiterstöbern auf beiden Seiten.

Hol dir agent-device

Kostenlos und offen — hol dir den Quellcode, gib dem Repository einen Stern und leg los.