Android emulator automation
Talos connects to Android through ADB. Give it a goal and it uses UI elements and screenshots to decide where to tap, what to type and when a visible checkpoint is complete.
Tap or long-press elements, swipe and scroll, enter or clear text, press Android navigation keys, open apps by package name, read screen text and wait for text to appear.
The observation → action → verification loop lets the agent choose a next step from the current interface. You do not have to specify every action in advance.
ADB connectivity is the starting point. Start your Android emulator and confirm it appears as an online device in adb devices. Physical Android phones also work when USB debugging is enabled and authorized.
There is no verified compatibility matrix for specific third-party emulator brands. App behavior, Android version and UI accessibility affect results. Talos does not control iOS simulators.
Your host. Your devices.
Run Talos on Windows or Linux and open its browser dashboard. View detected devices, assign a task to a device and inspect the current screen without starting a task.
ADB is required. FFmpeg is optional for live streaming. Configure a Claude API key or a custom HTTP provider that implements the Talos protocol. A generic chat endpoint is not automatically compatible.
View available Windows and Linux builds →
Use goal-driven interaction for app setup, repeated form entry and exploratory journeys where describing a result is easier than maintaining every UI action.
The agent can react to current labels and screen content, but this is not a guarantee that every dynamic UI will work. Test your app, keep the screen unlocked and use observable milestones.
Explore concrete Android workflow examples →| Your workflow | Scripted UI automation | Talos AI agent |
|---|---|---|
| Define the task | Write an action sequence and assertions. | Describe a goal and visible success criteria. |
| Respond to the screen | Use selectors and branches you maintain. | Choose the next action from the current UI and screenshots. |
| Investigate a run | Inspect the logs and captures your framework provides. | Review observations, actions, results, screenshots and token usage in one timeline. |
| Best fit | Exact assertions and repeatable regression gates. | Goal-driven app interaction, setup tasks and exploratory workflow checks. |
| Trade-off | Upfront scripting and maintenance. | Model latency, provider costs and decisions that can vary. |
AI decisions can vary and model calls add latency and cost. Talos is useful for visible interaction; it does not provide deterministic test assertions, a built-in CI test runner or guaranteed completion on every app.
The dashboard and run history stay on your host, while observations and relevant screenshots are sent to your selected model provider. Licensed installations contact the licence server, so offline operation is not promised.
Start with a single emulator and review the timeline. For continuous tasks, set a pause and limits for each cycle, and remember that costs accumulate across cycles.
Start with one workflow
Try it on your own emulator or phone. Bring an ADB connection, a clear goal and your model provider.
No card required for the licence trial. AI usage is billed separately.