### Start LocalAPIServer Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Example of how to start the shared instance of the LocalAPIServer.

```swift
LocalAPIServer.shared.start()
```

--------------------------------

### Swift Integration Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

Demonstrates starting and stopping the ASR service using Swift. Assumes shared instance access.

```swift
let asrService = AppServices.shared.asr
try await asrService.start()
let text = await asrService.stop()
```

--------------------------------

### Local API Usage Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Demonstrates how to start the Local API server and make a health check request from a client application.

```swift
// Start server
LocalAPIServer.shared.start()

// Client code (from any local app)
let url = URL(string: "http://127.0.0.1:47733/v1/health")!
let (data, response) = try await URLSession.shared.data(from: url)
let healthCheck = try JSONDecoder().decode(HealthResponse.self, from: data)
print("Status: \(healthCheck.status)")
```

--------------------------------

### Enable and Start Local API Server

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/errors.md

Configure user defaults to enable the local API and then start the LocalAPIServer. If the local API is disabled, the start() method will return silently.

```swift
// Configure to enable
UserDefaults.standard.set(true, forKey: "LocalAPIEnabled")
LocalAPIServer.shared.start()

// If disabled, start() returns silently

```

--------------------------------

### Client Libraries and Examples

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

The documentation covers examples for various client libraries including Python, JavaScript/Node.js, cURL, and Swift, along with categories like initialization, configuration, speech recognition, LLM requests, AI enhancement, local API usage, and error handling.

```APIDOC
## Supported Client Libraries and Example Categories

### Client Libraries
- Python (requests library)
- JavaScript/Node.js (fetch)
- cURL
- Swift

### Example Categories
- Initialization and setup
- Configuration
- Speech recognition
- LLM requests
- AI enhancement
- Local API usage
- Error handling
- Swift/Objective-C patterns
```

--------------------------------

### start()

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/asr-service.md

Initiates audio capture and prepares the ASR service for transcription. This method should be called before starting any transcription process.

```APIDOC
## Method: start()

### Description
Begin audio capture and prepare for transcription.

### Signature
```swift
func start() async throws
```

### Throws
Errors from microphone access or transcription provider preparation.

### Example
```swift
let asrService = ASRService()
do {
    try await asrService.start()
} catch {
    print("Failed to start ASR: \(error)")
}
```
```

--------------------------------

### Setup Pre-commit Hook

Source: https://github.com/altic-dev/fluidvoice/blob/main/README.md

Installs a pre-commit hook to prevent accidental commits of team IDs. Ensure you have execute permissions for the script.

```bash
cp scripts/check-team-id.sh .git/hooks/pre-commit
chmod +x .git/hooks/pre-commit
```

--------------------------------

### HotkeyShortcut displayString Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Demonstrates how to get a human-readable string representation of a HotkeyShortcut, including modifiers and the key name. Useful for displaying shortcuts to users.

```swift
let hotkey = HotkeyShortcut(keyCode: 49, modifierFlags: [.command])
print(hotkey.displayString)  // "⌘ + Space"
```

--------------------------------

### Initialize and Start ASR Service

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/asr-service.md

Initializes the ASRService and begins audio capture. Ensure microphone access is granted before starting.

```swift
let asrService = ASRService()
do {
    try await asrService.start()
} catch {
    print("Failed to start ASR: \(error)")
}
```

--------------------------------

### Start LocalAPIServer

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Starts the Local API server. It reads configuration, ensures it's enabled, and only accepts loopback connections.

```swift
func start() {
    // Implementation details omitted for brevity
}
```

--------------------------------

### LLM Request Example with Streaming

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/llm-client.md

Demonstrates how to configure and send an LLM request using the LLMClient. This example includes setting up messages, model, API key, and enabling streaming with a content chunk handler. It also shows how to catch and print LLM errors.

```swift
let config = LLMClient.Config(
    messages: [
        ["role": "system", "content": "You are a helpful assistant"],
        ["role": "user", "content": "Hello!"]
    ],
    model: "gpt-4",
    baseURL: "https://api.openai.com/v1",
    apiKey: "sk-...",
    streaming: true,
    tools: [],
    temperature: 0.7,
    onContentChunk: { chunk in
        print("Streaming: \(chunk)")
    }
)

do {
    let response = try await LLMClient.shared.request(config: config)
    print("Response: \(response.content)")
    if let thinking = response.thinking {
        print("Model thinking: \(thinking)")
    }
} catch let error as LLMError {
    print("LLM error: \(error.errorDescription ?? \"unknown\")")
}
```

--------------------------------

### Navigation and Entry Points

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

The documentation provides a clear navigation structure starting with README.md for an overview, INDEX.md for quick start and architecture, and specific service documentation within the api-reference/ directory.

```APIDOC
## Documentation Navigation Structure

### Entry Points
- **README.md**: Overview and document structure.
- **INDEX.md**: Architecture and common tasks.

### API Reference
- Located in the `api-reference/` directory.

### Suggested Reading Order
1. README.md
2. INDEX.md
3. types.md
4. configuration.md
5. Specific service docs (e.g., `asr-service.md`)
6. endpoints.md
7. errors.md
```

--------------------------------

### JavaScript/Node.js Client Examples

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Use the fetch API in JavaScript to communicate with the Fluidvoice API. Examples cover health checks, retrieving history, and transcribing audio.

```javascript
const BASE_URL = "http://127.0.0.1:47733/v1";

// Health check
fetch(`${BASE_URL}/health`)
  .then(r => r.json())
  .then(data => console.log(data));

// Get history
fetch(`${BASE_URL}/history?limit=10`)
  .then(r => r.json())
  .then(history => {
    history.forEach(item => {
      console.log(`${item.timestamp}: ${item.text}`);
    });
  });

// Transcribe audio
const audioBase64 = btoa(audioData);
fetch(`${BASE_URL}/transcribe`, {
  method: "POST",
  headers: {"Content-Type": "application/json"},
  body: JSON.stringify({audio: audioBase64, format: "wav"})
})
  .then(r => r.json())
  .then(result => console.log(result.text));
```

--------------------------------

### Swift Customization Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

Shows how to add a custom dictation prompt profile using Swift. Requires access to the shared SettingsStore.

```swift
let profile = SettingsStore.DictationPromptProfile(...)
SettingsStore.shared.addDictationPromptProfile(profile)
```

--------------------------------

### HotkeyShortcut Initialization and Usage Examples

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/types.md

Demonstrates how to create a HotkeyShortcut instance, display its string representation, check if an event matches the shortcut, and detect conflicts between modifier-only shortcuts.

```swift
// Create Command+Space
let hotkey = HotkeyShortcut(
    keyCode: 49,  // Space
    modifierFlags: [.command]
)
print(hotkey.displayString)  // "⌘ + Space"

// Check if event matches
let matches = hotkey.matches(keyCode: eventKeyCode, modifiers: eventModifiers)

// Detect conflicts between two modifier-only shortcuts
let conflict = shortcut1.conflictsWith(shortcut2)
```

--------------------------------

### Enable and Start Local API Server

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/INDEX.md

Enable the local API via UserDefaults and start the LocalAPIServer. Access the API at http://127.0.0.1:47733/v1/health.

```swift
UserDefaults.standard.set(true, forKey: "LocalAPIEnabled")
LocalAPIServer.shared.start()

// Then access at http://127.0.0.1:47733/v1/health
```

--------------------------------

### POST /v1/postprocess cURL Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Example of how to call the post-processing endpoint using cURL, specifying the text and mode.

```bash
curl -X POST http://127.0.0.1:47733/v1/postprocess \
  -H "Content-Type: application/json" \
  -d '{
    "text": "hello world",
    "mode": "dictate"
  }'
```

--------------------------------

### Install FluidVoice with Homebrew

Source: https://github.com/altic-dev/fluidvoice/blob/main/README.md

Use this command to install FluidVoice on your macOS system using Homebrew.

```bash
brew install --cask fluidvoice
```

--------------------------------

### Python Client Examples

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Interact with the Fluidvoice API using Python's requests library. Examples include health checks, fetching history, and adding dictionary replacements.

```python
import requests
import json

BASE_URL = "http://127.0.0.1:47733/v1"

# Health check
response = requests.get(f"{BASE_URL}/health")
print(response.json())

# Get history
response = requests.get(f"{BASE_URL}/history?limit=10")
history = response.json()
for item in history:
    print(f"{item['timestamp']}: {item['text']}")

# Add replacement
response = requests.post(
    f"{BASE_URL}/dictionary/replacements",
    json={"from": "ur", "to": "your"}
)
print(f"Status: {response.status_code}")
```

--------------------------------

### GET /v1/dictionary/replacements Response Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

The response is an array of objects, where each object specifies a word replacement from 'from' to 'to'.

```json
[
  {
    "from": "ur",
    "to": "your"
  },
  {
    "from": "teh",
    "to": "the"
  }
]
```

--------------------------------

### cURL Examples

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Perform API operations using cURL commands. Includes examples for health checks, fetching history with a limit, and adding custom words to the dictionary.

```bash
# Health check
curl http://127.0.0.1:47733/v1/health

# Get history
curl "http://127.0.0.1:47733/v1/history?limit=5"

# Add word to dictionary
curl -X POST http://127.0.0.1:47733/v1/dictionary/custom-words \
  -H "Content-Type: application/json" \
  -d '{"word":"MacBook","definition":"Apple laptop"}'
```

--------------------------------

### Example Usage of OpenAICompatibleProvider

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/ai-provider.md

Demonstrates how to use the OpenAICompatibleProvider to process text with an AI model. Ensure you have the provider initialized and provide valid API credentials and endpoint details.

```swift
let provider = OpenAICompatibleProvider()
let result = await provider.process(
    systemPrompt: "You are a helpful assistant",
    userText: "Fix the grammar: Im happy today",
    model: "gpt-4",
    apiKey: "sk-.",
    baseURL: "https://api.openai.com/v1",
    stream: false
)
print(result)  // "I'm happy today"
```

--------------------------------

### WhisperProvider Usage Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/transcription-provider.md

Demonstrates the typical workflow for using the WhisperProvider, including preparation, checking availability, capturing audio, transcribing, and clearing the cache. Ensure the provider is prepared before use.

```swift
// Create provider
let provider = WhisperProvider()

// Ensure ready
try await provider.prepare {
    updateUI(progress: progress)
}

// Check availability
guard provider.isAvailable && provider.isReady else {
    print("Provider not ready")
    return
}

// Transcribe samples
let samples = await captureAudio()
let result = try await provider.transcribe(samples)
print("You said: \(result.text)")

// Cleanup
try await provider.clearCache()
```

--------------------------------

### App Initialization with AppServices

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/app-services.md

Sets up the main application struct, injecting the AppServices singleton as an environment object. This pattern is used for top-level application setup.

```swift
@main
struct FluidApp: App {
    @StateObject private var appServices: AppServices
    
    init() {
        _appServices = StateObject(wrappedValue: AppServices.shared)
    }
    
    var body: some Scene {
        WindowGroup {
            ContentView()
                .environmentObject(appServices)
        }
    }
}
```

--------------------------------

### POST /v1/dictionary/replacements Response Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

A successful addition or update returns a 204 No Content status.

```http
HTTP/1.1 204 No Content
```

--------------------------------

### GET /v1/dictionary/replacements Request Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieve word replacement rules from the dictionary. The 'limit' parameter controls the number of results returned, with a maximum of 1000.

```http
GET /v1/dictionary/replacements?limit=100 HTTP/1.1
Host: 127.0.0.1:47733
```

--------------------------------

### Stop LocalAPIServer Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Example of how to stop the shared instance of the LocalAPIServer.

```swift
LocalAPIServer.shared.stop()
```

--------------------------------

### Ensure Transcription Provider is Ready with Progress

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/asr-service.md

Downloads and prepares the necessary speech recognition model. Provides a progress handler to monitor download status. This method should be called before starting transcription.

```swift
do {
    try await asrService.ensureTranscriptionProviderReady { progress in
        print("Download: \(Int(progress * 100))%")
    }
} catch {
    print("Model preparation failed: \(error)")
}
```

--------------------------------

### POST /v1/postprocess Response Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

The response contains the enhanced text and a boolean indicating if any modifications were made.

```json
{
  "text": "Hello world.",
  "changed": true
}
```

--------------------------------

### getAppPromptBinding

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Get the prompt binding configuration for a specific application identified by its bundle ID. Returns the binding details or nil if none is configured.

```APIDOC
## getAppPromptBinding(forBundleID:)

### Description
Get the prompt binding for a specific app.

### Method
func getAppPromptBinding(forBundleID bundleID: String) -> AppPromptBinding?

### Parameters
#### Path Parameters
- **bundleID** (String) - Required - Bundle ID of the app to retrieve binding for

### Returns
Binding or nil if not configured
```

--------------------------------

### Integrate ASRService with Dictation Post-Processing

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Shows how to start and stop the ASRService, which automatically utilizes the DictationPostProcessingService for enhanced transcriptions if enabled.

```swift
let asrService = ASRService()
try await asrService.start()
// ... speak ...
let transcribed = await asrService.stop()
// If enhancement is enabled, finalText is already enhanced
// Otherwise it's raw transcription
```

--------------------------------

### POST /v1/dictionary/replacements Request Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Add or update a word replacement rule in the dictionary. Both 'from' and 'to' fields are required.

```json
{
  "from": "ur",
  "to": "your"
}
```

--------------------------------

### Generic Error Response

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Example JSON format for all error responses, such as route not found.

```json
{
  "error": "Route not found."
}
```

--------------------------------

### API Documentation Structure

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

Each API document includes the exact signature from the source, parameter tables with types and defaults, return types and descriptions, error conditions, and minimal examples. Cross-references and hierarchical navigation are also provided.

```APIDOC
## API Document Contents

### Signature
Exact signature from source code.

### Parameters
- **paramName** (type) - Required/Optional - Description with default values if applicable.

### Returns
- **type** - Description of the return value.

### Errors
- **ErrorType** - Description of when this error is thrown or rejected.

### Examples
Minimal but complete code examples demonstrating usage.

### Source Reference
- **File**: path/to/source/file.ext
- **Line**: line_number
```

--------------------------------

### POST /v1/postprocess Request Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Send transcribed text to the post-processing endpoint for AI-enhanced output. The 'mode' parameter can be set to 'dictate' or 'edit', and 'context' provides surrounding text for better results.

```json
{
  "text": "hello world",
  "mode": "dictate",
  "context": "Selected text before dictation"
}
```

--------------------------------

### Retrieve All Dictation Prompt Profiles

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Gets a list of all custom dictation prompt profiles currently configured. Returns an empty array if none are set.

```swift
func allDictationPromptProfiles() -> [DictationPromptProfile]
```

--------------------------------

### GET /v1/history Request

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Example of a GET request to retrieve transcription history with pagination parameters.

```http
GET /v1/history?limit=50&offset=0 HTTP/1.1
Host: 127.0.0.1:47733
```

--------------------------------

### HotkeyShortcut isModifierOnlyShortcut Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Shows how to determine if a HotkeyShortcut is triggered by a modifier key alone. This is useful for differentiating between shortcuts that require a modifier and another key, versus those that are just the modifier itself.

```swift
// Command+V (regular key)
let regular = HotkeyShortcut(keyCode: 9, modifierFlags: [.command])
print(regular.isModifierOnlyShortcut)  // false

// Right Command (modifier only)
let modifier = HotkeyShortcut(keyCode: 54, modifierFlags: [])
print(modifier.isModifierOnlyShortcut)  // true
```

--------------------------------

### GET /v1/history Response (200 OK)

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Example JSON response for a successful retrieval of transcription history.

```json
[
  {
    "id": "abc123",
    "text": "hello world",
    "timestamp": "2025-12-21T10:30:00Z",
    "model": "parakeet-tdt-v3",
    "duration_ms": 2500
  }
]
```

--------------------------------

### Basic Dictation with ASR Service

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Starts the ASR service, captures audio input, and stops the service to retrieve transcribed text. The output will be enhanced text if AI is enabled, otherwise raw transcription.

```swift
let asrService = AppServices.shared.asr
try await asrService.start()
// ... speak ...
let text = await asrService.stop()
print(text)  // Enhanced text if AI enabled, raw otherwise
```

--------------------------------

### Startup Flow Visualization

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Illustrates the sequence of events during application startup, highlighting lazy service initialization after the UI is ready to prevent crashes.

```text
1. AppServices singleton created (services not yet initialized)
   ↓
2. UI renders and completes (1.5s delay for AttributeGraph)
   ↓
3. signalUIReady() called
   ↓
4. initializeServicesIfNeeded() called
   ↓
5. Services lazily initialized on first access
```

--------------------------------

### GET /v1/dictionary/custom-words Response Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

The response is an array of custom word objects, each containing the 'word' and an optional 'definition'.

```json
[
  {
    "word": "Kubernetes",
    "definition": "container orchestration platform"
  }
]
```

--------------------------------

### GET /v1/dictionary/custom-words Request Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieve custom words from the dictionary. The 'limit' parameter restricts the number of words returned, up to a maximum of 1000.

```http
GET /v1/dictionary/custom-words?limit=100 HTTP/1.1
Host: 127.0.0.1:47733
```

--------------------------------

### prepare(progressHandler:)

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/transcription-provider.md

Downloads and/or loads transcription models. It can optionally report download progress.

```APIDOC
## prepare(progressHandler:)

### Description
Download and/or load models for transcription. This method can optionally provide progress updates during the download process.

### Method
`async throws`

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Parameters
- **progressHandler** ((Double) -> Void)? - No - Callback for download progress (0.0 to 1.0)

### Throws
Error if download or model loading fails

### Example
```swift
let provider = ParakeetRealtimeProvider()
do {
    try await provider.prepare {
        progress in print("Download: \(Int(progress * 100))%")
    }
} catch {
    print("Failed to prepare: \(error)")
}
```
```

--------------------------------

### Prepare Transcription Models

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/transcription-provider.md

Downloads and loads transcription models. Use the progress handler to monitor download progress. This method can throw an error if preparation fails.

```swift
let provider = ParakeetRealtimeProvider()
do {
    try await provider.prepare {
        progress in
        print("Download: \(Int(progress * 100))%")
    }
} catch {
    print("Failed to prepare: \(error)")
}
```

--------------------------------

### Creating and Using Custom Prompt Profiles

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Illustrates how to define a custom `DictationPromptProfile` for specific formatting (e.g., code comments) and add it to the `SettingsStore`. The custom profile can then be used for text enhancement.

```swift
let codePrompt = SettingsStore.DictationPromptProfile(
    name: "Code Comments",
    prompt: "Format as a concise code comment preserving technical terms",
    mode: .edit,
    includeContext: false
)
SettingsStore.shared.addDictationPromptProfile(codePrompt)

// Later, use it
let dictated = "creates a new user with email and password"
let enhanced = await service.process(
    text: dictated,
    mode: .edit,
    context: nil
)
// "Creates a new user with email and password."
```

--------------------------------

### Health Check Endpoint

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Example of how to perform a health check on the Local API.

```APIDOC
## Usage Example

```swift
// Start server
LocalAPIServer.shared.start()

// Client code (from any local app)
let url = URL(string: "http://127.0.0.1:47733/v1/health")!
let (data, response) = try await URLSession.shared.data(from: url)
let healthCheck = try JSONDecoder().decode(HealthResponse.self, from: data)
print("Status: \(healthCheck.status)")
```

### GET /v1/health

#### Description
Checks the health status of the Local API.

#### Endpoint
`/v1/health`

#### Response
##### Success Response (200)
- **status** (string) - The health status of the API (e.g., "ok").
```

--------------------------------

### Enable Local API

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Configuration instructions for enabling the Local API via UserDefaults or the Settings UI.

```APIDOC
## Configuration

Enable Local API in UserDefaults:

```swift
let defaults = UserDefaults.standard
defaults.set(true, forKey: "LocalAPIEnabled")
defaults.set(47733, forKey: "LocalAPIPort")
```

Or via Settings UI:
`Settings → Advanced → Local API → Enable`
```

--------------------------------

### Get History

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieves a list of historical transcription data, with an optional limit.

```APIDOC
## GET /history

### Description
Retrieves a list of historical transcription data.

### Method
GET

### Endpoint
/v1/history

### Parameters
#### Query Parameters
- **limit** (integer) - Optional - The maximum number of history items to retrieve.
```

--------------------------------

### Configure Hotkey and Local API

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Defines a command + space hotkey for dictation and enables the local API server.

```swift
// Hotkey
let hotkey = HotkeyShortcut(keyCode: 49, modifierFlags: [.command])
SettingsStore.shared.dictationHotkey = hotkey

// Local API
UserDefaults.standard.set(true, forKey: "LocalAPIEnabled")
LocalAPIServer.shared.start()
```

--------------------------------

### Start and Stop Dictation

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/INDEX.md

Initiate speech-to-text dictation and retrieve the transcribed text upon stopping.

```swift
let asrService = AppServices.shared.asr
try await asrService.start()
// ... speak ...
let text = await asrService.stop()
```

--------------------------------

### Process Text with O1 Reasoning Model

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/ai-provider.md

Engage a specific reasoning model like 'o1' using the OpenAICompatibleProvider. This model typically does not use temperature settings.

```swift
let provider = OpenAICompatibleProvider()
let result = await provider.process(
    systemPrompt: "You are a problem solver.",
    userText: "Find all integer solutions to x^2 + y^2 = 25",
    model: "o1",  // Reasoning model (no temperature)
    apiKey: "sk-...",
    baseURL: "https://api.openai.com/v1",
    stream: false
)
print(result)
```

--------------------------------

### GET /v1/dictionary/custom-words

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieves custom words from the dictionary. Supports limiting the number of words returned.

```APIDOC
## GET /v1/dictionary/custom-words

### Description
Get custom dictionary. This endpoint retrieves a list of custom words added to the dictionary, optionally including their definitions.

### Method
GET

### Endpoint
/v1/dictionary/custom-words

### Parameters
#### Query Parameters
- **limit** (integer) - No - Default: 100, Maximum: 1000 - Number of words to return

### Response
#### Success Response (200 OK)
- **Array of word objects**
  - **word** (string) - Word
  - **definition** (string) - Definition (optional)

#### Response Example
```json
[
  {
    "word": "Kubernetes",
    "definition": "container orchestration platform"
  }
]
```
```

--------------------------------

### Convert Key Code to String with Keyboard Layout Support

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Illustrates converting a key code to its string representation, automatically handling different keyboard layouts. Falls back to QWERTY if the current layout is not supported.

```swift
// On French AZERTY: '2' might be displayed differently
let qwerty = HotkeyShortcut.keyCodeToString(19)  // "2" (QWERTY)
// On AZERTY: might be "é" or different character
```

--------------------------------

### Get Transcription History

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieves a paginated list of past transcriptions. Supports filtering by limit and offset.

```APIDOC
## GET /v1/history

### Description
Retrieve transcription history.

### Method
GET

### Endpoint
/v1/history

### Query Parameters
- **limit** (integer) - Optional - Number of items to return. Default: 100, Maximum: 1000
- **offset** (integer) - Optional - Pagination offset. Default: 0

### Request Example
```http
GET /v1/history?limit=50&offset=0 HTTP/1.1
Host: 127.0.0.1:47733
```

### Response
#### Success Response (200 OK)
- **id** (string) - Unique identifier
- **text** (string) - Transcribed text
- **timestamp** (string) - When transcribed (ISO8601 format)
- **model** (string) - ASR model used
- **duration_ms** (integer) - Recording duration in milliseconds

#### Response Example
```json
[
  {
    "id": "abc123",
    "text": "hello world",
    "timestamp": "2025-12-21T10:30:00Z",
    "model": "parakeet-tdt-v3",
    "duration_ms": 2500
  }
]
```

### Errors
- `500` — History lookup failed
```

--------------------------------

### createAppPromptBinding

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Create a binding between an application and a prompt profile. This allows specific apps to use custom prompts.

```APIDOC
## createAppPromptBinding(mode:appBundleID:appName:promptID:)

### Description
Create a binding between an app and a prompt profile.

### Method
func createAppPromptBinding(
    mode: PromptMode,
    appBundleID: String,
    appName: String,
    promptID: String?
)

### Parameters
#### Path Parameters
- **mode** (PromptMode) - Required - Dictate or edit mode
- **appBundleID** (String) - Required - Bundle identifier of target app
- **appName** (String) - Required - Human-readable app name
- **promptID** (String?) - Optional - Profile ID or nil for default
```

--------------------------------

### GET /v1/dictionary/replacements

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Retrieves a list of word replacement rules. Supports limiting the number of results returned.

```APIDOC
## GET /v1/dictionary/replacements

### Description
Get word replacements. This endpoint retrieves a list of predefined word replacements that can be used to automatically correct or standardize text.

### Method
GET

### Endpoint
/v1/dictionary/replacements

### Parameters
#### Query Parameters
- **limit** (integer) - No - Default: 100, Maximum: 1000 - Number of replacements to return

### Response
#### Success Response (200 OK)
- **Array of replacement objects**
  - **from** (string) - Text to replace
  - **to** (string) - Replacement text

#### Response Example
```json
[
  {
    "from": "ur",
    "to": "your"
  },
  {
    "from": "teh",
    "to": "the"
  }
]
```
```

--------------------------------

### SettingsStore.DictationPromptProfile

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/types.md

Defines a template for AI enhancement prompts used in dictation settings.

```APIDOC
## SettingsStore.DictationPromptProfile

### Description
Template for AI enhancement prompts.

### Properties
- **id** (String) - Unique identifier (UUID)
- **name** (String) - Display name
- **prompt** (String) - Prompt template text
- **mode** (PromptMode) - Dictate or edit mode
- **includeContext** (Bool) - Include selected text as context
- **createdAt** (Date) - Creation timestamp
- **updatedAt** (Date) - Last modification timestamp
```

--------------------------------

### POST /v1/dictionary/custom-words Request Example

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Add a word to the custom dictionary. The 'word' is mandatory, and a 'definition' can be optionally provided.

```json
{
  "word": "Kubernetes",
  "definition": "container orchestration platform"
}
```

--------------------------------

### allDictationPromptProfiles

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Get all custom prompt profiles currently stored. This method returns an array containing all available profiles.

```APIDOC
## allDictationPromptProfiles()

### Description
Get all custom prompt profiles.

### Method
func allDictationPromptProfiles() -> [DictationPromptProfile]

### Returns
Array of all profiles
```

--------------------------------

### init(keyCode:modifierFlags:modifierKeyCodes:)

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Initializes a HotkeyShortcut with a primary key code and modifier information. It supports automatic validation and normalization of modifiers, with an option to provide explicit modifier key codes.

```APIDOC
## init(keyCode:modifierFlags:modifierKeyCodes:)

### Description
Creates a hotkey with automatic validation and normalization.

### Parameters
#### Path Parameters
- **keyCode** (UInt16) - Required - Primary key code
- **modifierFlags** (NSEvent.ModifierFlags) - Required - Modifier flags
- **modifierKeyCodes** ([UInt16]) - Optional - Explicit modifier key codes (defaults to [])

### Details
- Normalizes modifier key codes (deduplicates and sorts).
- Validates modifier flags.
- If modifierKeyCodes is provided, uses them instead of modifierFlags.
- Removes trigger key from combined modifiers.

### Request Example
```swift
// Command+Space
let hotkey = HotkeyShortcut(
    keyCode: 49,  // Space
    modifierFlags: [.command]
)

// Or with explicit modifier key codes
let hotkey2 = HotkeyShortcut(
    keyCode: 49,
    modifierFlags: [],
    modifierKeyCodes: [55, 49]  // Command + Space
)
```
```

--------------------------------

### App-Specific Enhancement Configuration

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Details how to create an app-specific prompt profile and bind it to a particular application using its bundle ID. This allows for tailored dictation enhancement based on the active application.

```swift
// Configure different enhancement per app
let scriptPrompt = SettingsStore.DictationPromptProfile(
    name: "AppleScript",
    prompt: "Format as AppleScript code, preserve technical syntax",
    mode: .edit,
    includeContext: false
)
SettingsStore.shared.addDictationPromptProfile(scriptPrompt)

SettingsStore.shared.createAppPromptBinding(
    mode: .edit,
    appBundleID: "com.apple.ScriptEditor2",
    appName: "Script Editor",
    promptID: scriptPrompt.id
)

// When dictating in Script Editor, uses ScriptPrompt
let enhanced = await service.process(
    text: "delay for one second",
    mode: .edit,
    context: nil
)
// "delay for 1 second" (preserves AppleScript style)
```

--------------------------------

### Run Unit Tests

Source: https://github.com/altic-dev/fluidvoice/blob/main/docs/MACOS_UI_AUTOMATION_BRANCH_PLAN.md

Command to execute unit tests for the Fluid project. Ensure the correct scheme and destination are specified.

```bash
xcodebuild test -project Fluid.xcodeproj -scheme Fluid -destination 'platform=macOS,arch=arm64' -only-testing:FluidTests
```

--------------------------------

### Clone FluidVoice Repository and Open Project

Source: https://github.com/altic-dev/fluidvoice/blob/main/README.md

Clone the FluidVoice repository from GitHub and open the Xcode project. Dependencies are managed via Swift Package Manager.

```bash
git clone https://github.com/altic-dev/FluidVoice.git
cd FluidVoice
open Fluid.xcodeproj
```

--------------------------------

### Enable Local API in UserDefaults

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/local-api.md

Configure the Local API by setting 'LocalAPIEnabled' to true and specifying the port in UserDefaults.

```swift
let defaults = UserDefaults.standard
defaults.set(true, forKey: "LocalAPIEnabled")
defaults.set(47733, forKey: "LocalAPIPort")
```

--------------------------------

### Local API Health Check

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Performs a health check on the local FluidVoice API. This is a simple GET request to the health endpoint.

```bash
# Health check
curl http://127.0.0.1:47733/v1/health
```

--------------------------------

### GET /v1/health

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/endpoints.md

Checks the health of the FluidVoice server and retrieves its current version. This endpoint is useful for verifying that the Local API is running and accessible.

```APIDOC
## GET /v1/health

### Description
Server health check and version info.

### Method
GET

### Endpoint
/v1/health

### Response
#### Success Response (200 OK)
- **status** (string) - Server status ("ok")
- **version** (string) - App version

#### Response Example
```json
{
  "status": "ok",
  "version": "1.6.0"
}
```
```

--------------------------------

### Check and Load ASR Model Readiness

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/errors.md

Before transcribing, verify if the Automatic Speech Recognition (ASR) service is ready. If models exist on disk, load them; otherwise, initiate a download.

```swift
// Check before transcribing
if !asrService.isAsrReady {
    if asrService.modelsExistOnDisk {
        // Load model
        try await asrService.ensureTranscriptionProviderReady()
    } else {
        // Download model
        try await asrService.ensureTranscriptionProviderReady {
            progress in
            print("Download: \(Int(progress * 100))%")
        }
    }
}
```

--------------------------------

### SettingsStore Class Signature

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Defines the SettingsStore class, its singleton instance, and constants for transcription preview character limits.

```swift
final class SettingsStore: ObservableObject {
    static let shared = SettingsStore()
    static let transcriptionPreviewCharLimitRange: ClosedRange<Int> = 50...800
    static let transcriptionPreviewCharLimitStep = 50
    static let defaultTranscriptionPreviewCharLimit = 150
}
```

--------------------------------

### Get Modifier Flag for Key Code

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Retrieves the NSEvent.ModifierFlags corresponding to a given modifier key code. Provides a mapping for common modifier keys.

```swift
static func modifierFlag(forKeyCode keyCode: UInt16) -> NSEvent.ModifierFlags?
```

```swift
let flag = HotkeyShortcut.modifierFlag(forKeyCode: 55)
print(flag)  // Optional(.command)
```

--------------------------------

### Get AI Enhancement Mode

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Retrieve the current AI enhancement mode for dictation. Modes range from off to default, private AI, or a custom profile.

```swift
SettingsStore.shared.dictationAIMode
```

--------------------------------

### Access Audio Hardware Observer

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/app-services.md

Access the audio hardware observer. It is lazily created upon first access, not during AppServices initialization.

```swift
let observer = appServices.audioObserver
print("Microphone: \(observer.defaultMicrophoneDevice)")
```

--------------------------------

### OpenAICompatibleProvider.process

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/ai-provider.md

Sends text to an OpenAI-compatible API and returns the response. It handles various configurations like streaming, local endpoints, and specific model requirements.

```APIDOC
## func process(systemPrompt:userText:model:apiKey:baseURL:stream:)

### Description
Sends text to an OpenAI-compatible API and returns the response. It handles various configurations like streaming, local endpoints, and specific model requirements.

### Method
`async func`

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Parameters
- **systemPrompt** (String) - Required - System instruction
- **userText** (String) - Required - User message
- **model** (String) - Required - Model identifier
- **apiKey** (String) - Required - API authentication key
- **baseURL** (String) - Required - API base URL
- **stream** (Bool) - Optional - Default: `false` - Enable streaming

### Returns
Content from first response choice, or error message

### Error Handling:
- Invalid URL → "Error: Invalid Base URL"
- HTTP errors → "Error: HTTP {code}: {body}"
- JSON encoding failure → "Error: Failed to encode request"
- Network/connection errors → "Error: {error description}"
- Decoding errors → "Error: {error description}"

### Special Behavior:
1. **URL Handling:** If baseURL doesn't contain /chat/completions, /api/chat, or /api/generate, automatically appends /chat/completions
2. **Local Endpoints:** For loopback addresses (127.0.0.1, localhost, 10.x.x.x, etc.), omits Authorization header
3. **Temperature Handling:** For reasoning models (o1, o3, gpt-5), omits temperature parameter
4. **Reasoning Effort:** For gpt-oss models (Groq), adds `reasoning_effort: "low"`
5. **Default Temperature:** Sets temperature to 0.2 for non-reasoning models
```

--------------------------------

### audioObserver

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/app-services.md

Provides access to the audio hardware observer. This property is lazily initialized, meaning the observer is only created when it is first accessed.

```APIDOC
## audioObserver

### Description
Get or lazily create the audio hardware observer.

### Method
var

### Parameters
None

### Returns
AudioHardwareObserver instance (created on first access)

### Details
- The observer is only created when first accessed, not at AppServices initialization.

### Example
```swift
let observer = appServices.audioObserver
print("Microphone: \(observer.defaultMicrophoneDevice)")
```
```

--------------------------------

### Compare HotkeyShortcut Instances

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Shows how to compare two HotkeyShortcut instances for equality. Equality is determined by matching modifier key codes or, if normalized, by matching keyCode and relevantModifierFlags.

```swift
let hotkey1 = HotkeyShortcut(keyCode: 49, modifierFlags: [.command])
let hotkey2 = HotkeyShortcut(keyCode: 49, modifierFlags: [.command, .option])

hotkey1 == hotkey1  // true
hotkey1 == hotkey2  // false
```

--------------------------------

### Store and Retrieve API Key

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Demonstrates how to store and retrieve an API key using the SettingsStore. The key is automatically encrypted and decrypted.

```swift
// Store a key
SettingsStore.shared.openAIKey = "sk-..."

// Retrieve (automatically decrypted)
if let key = SettingsStore.shared.openAIKey {
    // use key
}
```

--------------------------------

### Initialize HotkeyShortcut with Modifiers

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/hotkey-shortcut.md

Create a hotkey using primary key code and modifier flags. Supports explicit modifier key codes as an alternative.

```swift
init(
    keyCode: UInt16,
    modifierFlags: NSEvent.ModifierFlags,
    modifierKeyCodes: [UInt16] = []
)
```

```swift
// Command+Space
let hotkey = HotkeyShortcut(
    keyCode: 49,  // Space
    modifierFlags: [.command]
)

// Or with explicit modifier key codes
let hotkey2 = HotkeyShortcut(
    keyCode: 49,
    modifierFlags: [],
    modifierKeyCodes: [55, 49]  // Command + Space
)
```

--------------------------------

### Get App Prompt Binding for Bundle ID

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Retrieves the prompt binding configuration for a given application bundle identifier. Returns `nil` if no binding is configured for the app.

```swift
func getAppPromptBinding(forBundleID bundleID: String) -> AppPromptBinding?
```

--------------------------------

### Process Text with Local LM Studio

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/ai-provider.md

Connect to a locally running language model via LM Studio using the OpenAICompatibleProvider. The API key is not used for local endpoints.

```swift
let provider = OpenAICompatibleProvider()
let result = await provider.process(
    systemPrompt: "You are helpful.",
    userText: "What is 2+2?",
    model: "local-model",
    apiKey: "",  // Not used for local endpoints
    baseURL: "http://localhost:1234/v1",
    stream: false
)
print(result)  // "2+2 equals 4"
```

--------------------------------

### Create App Prompt Binding

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/settings-store.md

Establishes a link between an application and a prompt profile. You can specify a particular prompt profile or use the default.

```swift
func createAppPromptBinding(
    mode: PromptMode,
    appBundleID: String,
    appName: String,
    promptID: String?
)
```

--------------------------------

### Local API Response Constructor

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/types.md

Initializes a Local API response. Defaults are provided for headers and body.

```swift
init(status: Int, headers: [String: String] = [:], body: Data = Data())
```

--------------------------------

### Local API Get Transcription History

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Retrieves a limited number of transcription history records from the local FluidVoice API. The 'limit' query parameter controls the number of results returned.

```bash
# Get history
curl http://127.0.0.1:47733/v1/history?limit=10
```

--------------------------------

### initializeServicesIfNeeded()

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/app-services.md

Safely initializes all services after the UI is ready. This method is idempotent and can be called multiple times, but services will only be initialized once.

```APIDOC
## initializeServicesIfNeeded()

### Description
Safely initialize all services after the UI is ready.

### Method
func

### Parameters
None

### Details
- Call this from ContentView.onAppear after signalUIReady()
- Safe to call multiple times—only initializes once
- Logs initialization progress

### Example
```swift
.onAppear {
    Task {
        try? await Task.sleep(nanoseconds: 1_500_000_000)
        appServices.signalUIReady()
        appServices.initializeServicesIfNeeded()
    }
}
```
```

--------------------------------

### Get Selected AI Provider

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Retrieve the currently selected AI provider for dictation processing. Options include OpenAI, Groq, Anthropic, Cohere, custom endpoints, and local models.

```swift
SettingsStore.shared.selectedAIProvider
```

--------------------------------

### Documentation Verification and Quality Checks

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/GENERATION_SUMMARY.txt

The documentation undergoes rigorous quality checks to ensure completeness, accuracy, and adherence to technical precision, including verifying all public elements, tested examples, and consistent formatting.

```APIDOC
## Documentation Quality Assurance

### Completeness Checks
- All public classes and structs documented
- All public methods documented
- All public properties documented
- All configuration options documented
- All HTTP endpoints documented
- Error handling documented
- Thread safety notes included

### Quality Checks
- No placeholder content
- All examples tested against source code
- All signatures match source code
- All links are internal and valid
- Consistent formatting throughout
- No marketing copy or introductions
- Focus on technical precision
```

--------------------------------

### Configure SettingsStore in Swift

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/configuration.md

Use SettingsStore.shared to access and modify application settings. This includes setting speech models, enabling AI enhancements with API keys, defining hotkeys, creating custom prompt profiles, and binding profiles to specific applications. It also covers enabling and configuring local audio history.

```swift
let settings = SettingsStore.shared

// Configure speech model
settings.selectedSpeechModel = .parakeetTDT

// Enable AI enhancement with OpenAI
settings.selectedAIProvider = .openai
settings.openAIKey = "sk-..."
settings.aiEnhancementEnabled = true

// Set dictation hotkey (Command+Space)
let hotkey = HotkeyShortcut(
    keyCode: 49,
    modifierFlags: [.command]
)
settings.dictationHotkey = hotkey

// Create a custom prompt profile
let profile = SettingsStore.DictationPromptProfile(
    name: "Email",
    prompt: "Format as professional email",
    mode: .edit,
    includeContext: true
)
settings.addDictationPromptProfile(profile)

// Bind profile to Mail app
settings.createAppPromptBinding(
    mode: .edit,
    appBundleID: "com.apple.mail",
    appName: "Mail",
    promptID: profile.id
)

// Enable local audio history
settings.recordAudioHistory = true
settings.audioHistoryBudgetMB = 1000
```

--------------------------------

### Get Selected Prompt Profile

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Retrieves the active prompt profile for a given app bundle ID and processing mode. Falls back to the default profile if no app-specific binding is found. Returns nil if the mode is set to .off.

```swift
let prompt = service.getSelectedPrompt(
    for: "com.apple.mail",
    mode: .edit
)
if let prompt = prompt {
    print("Using profile: \(prompt.name)")
    print("Prompt: \(prompt.prompt)")
}
```

--------------------------------

### Create a Custom Dictation Prompt Profile

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/INDEX.md

Define a custom prompt profile for dictation, specifying its name, prompt text, mode, and context inclusion.

```swift
let profile = SettingsStore.DictationPromptProfile(
    name: "Email",
    prompt: "Format as professional email",
    mode: .edit,
    includeContext: true
)
SettingsStore.shared.addDictationPromptProfile(profile)
```

--------------------------------

### Create Custom Prompt Profile

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/api-reference/dictation-post-processing.md

Define and add a custom prompt profile for dictation. This allows for specific formatting and enhancement instructions, including context inclusion.

```swift
let profile = SettingsStore.DictationPromptProfile(
    name: "Email",
    prompt: "Format as a professional email with proper grammar",
    mode: .edit,
    includeContext: true
)
SettingsStore.shared.addDictationPromptProfile(profile)
```

--------------------------------

### Configure Speech and AI Settings

Source: https://github.com/altic-dev/fluidvoice/blob/main/_autodocs/README.md

Sets the default speech recognition model to Parakeet TDT and enables OpenAI for AI enhancement, requiring an API key.

```swift
// Speech model
SettingsStore.shared.selectedSpeechModel = .parakeetTDT

// AI enhancement
SettingsStore.shared.selectedAIProvider = .openai
SettingsStore.shared.openAIKey = "sk-..."
SettingsStore.shared.aiEnhancementEnabled = true
```