# Automated iOS App Testing

Combine Synapse's memory system with the Computer Control API to build
LLM-driven iOS app tests. The LLM remembers test scenarios, learns from past
failures, and adapts to UI changes.

## Architecture

```
┌──────────────┐    commands    ┌──────────────┐    screenshots    ┌──────────────┐
│  LLM Agent   │ ─────────────▶│  Synapse     │ ────────────────▶ │  iOS Sim     │
│  (Claude)    │               │  Computer    │ ◀──────────────── │  (via agent) │
└──────────────┘               │  Control     │    results        └──────────────┘
       │                       └──────────────┘
       │ store/recall
       ▼
┌──────────────┐
│  Memories    │ (test scenarios, past failures, UI patterns)
└──────────────┘
```

## Prerequisites

- Synapse account + Mind Key
- Synapse MCP server configured in Claude Desktop
- iOS Simulator with `screen-remote-agent` installed
- Computer registered in Synapse (see [Computer Control API](/docs/api/computers))

## Step 1: Register the Simulator Computer

On the Mac running iOS Simulator:

```bash
# Get install code from Synapse
curl -X POST https://synapse.schaefer.zone/computers/install-code \
  -H "Authorization: Bearer YOUR_MIND_KEY" \
  -d '{"computer_name":"ios-sim"}'
# → { "install_code": "ic_..." }

# Run screen-remote-agent on the Mac
# (uses the install code to register)
```

## Step 2: Store Test Scenarios in Memory

Store reusable test scenarios as memories:

```python
import requests

def store_test_scenario(name, steps, app):
    requests.post(f"{URL}/memory",
        headers={"Authorization": f"Bearer {MIND_KEY}"},
        json={
            "category": "skill",
            "key": f"test_scenario_{name}",
            "content": f"App: {app}\nSteps:\n" + "\n".join(steps),
            "tags": ["test", "ios", app],
            "priority": "high"
        })

store_test_scenario("login_flow", [
    "Launch app",
    "Tap email field",
    "Type test@example.com",
    "Tap password field",
    "Type password123",
    "Tap Login button",
    "Verify home screen appears"
], "MyApp")
```

## Step 3: LLM-Driven Test Execution

In Claude Desktop (with Synapse MCP configured):

```
Run the login_flow test scenario on the iOS Simulator.
Take a screenshot after each step and verify the expected UI.
If any step fails, store the failure as a memory so we can
avoid it next time.
```

Claude will:

1. Call `memory_search` to find the `test_scenario_login_flow` memory
2. Call `computer_screenshot` to see the current state
3. Execute each step via `computer_command_queue` (click, type)
4. Verify results via screenshots
5. Store any failures as `mistake` memories

## Step 4: Self-Healing Tests

When a test fails, store the failure and the recovery:

```python
def store_test_failure(scenario, step, error, recovery):
    requests.post(f"{URL}/memory",
        headers={"Authorization": f"Bearer {MIND_KEY}"},
        json={
            "category": "mistake",
            "key": f"failure_{scenario}_{step}",
            "content": f"Scenario: {scenario}\nStep: {step}\nError: {error}\nRecovery: {recovery}",
            "tags": ["test", "failure", "ios", scenario],
            "priority": "high"
        })

# Example
store_test_failure("login_flow", "tap_login",
    "Login button not found at expected coordinates",
    "Button moved due to new logo. Search by accessibility label instead.")
```

Next time the LLM runs the test, it recalls the failure and applies the
recovery automatically.

## Step 5: Test Result Tracking

Track test runs as tasks:

```python
def track_test_run(scenario, status, duration):
    requests.post(f"{URL}/mind/task",
        headers={"Authorization": f"Bearer {MIND_KEY}",
                 "Content-Type": "application/json"},
        json={
            "title": f"Test: {scenario}",
            "description": f"Status: {status}, Duration: {duration}s",
            "priority": "normal"
        })
```

## Common Commands

| Action | Command |
|--------|---------|
| Launch Simulator | `xcrun simctl launch booted com.example.app` |
| Screenshot | `computer_screenshot` (via Synapse MCP) |
| Tap at (x,y) | `computer_command_queue {type:"click", payload:{x,y}}` |
| Type text | `computer_command_queue {type:"type", payload:{text:"..."}}` |
| Press Home | `computer_command_queue {type:"key", payload:{keys:["Cmd","Shift","H"]}}` |

## Best Practices

> [!TIP]
> - **Store UI coordinates as memories** — UI changes, but the LLM can re-learn
> - **Use accessibility labels** — more stable than coordinates
> - **Store test data separately** — use variables for usernames, passwords
> - **Run tests in clean state** — reset Simulator between runs
> - **Log screenshots for failures** — useful for debugging

## Next Steps

- [Self-Healing Tests](/docs/guides/self-healing-tests)
- [Computer Control API](/docs/api/computers)
- [Memory Best Practices](/docs/guides/memory-best-practices)