Murf TTS service documentation

mint.json

@@ -221,6 +221,7 @@
             "server/services/tts/groq",
             "server/services/tts/lmnt",
             "server/services/tts/minimax",
+            "server/services/tts/murf",
             "server/services/tts/neuphonic",
             "server/services/tts/riva",
             "server/services/tts/openai",

server/services/tts/murf.mdx

@@ -0,0 +1,162 @@
+---
+title: "Murf.ai"
+description: "Text-to-speech services using Murf's WebSocket-based streaming API"
+---
+
+## Overview
+
+Murf’s API delivers the most scalable and reliable AI voice generation with **99.38% pronunciation accuracy**. It gives developers access to **150+ natural-sounding voices** across **21 languages** and **20+ speaking styles**, ranging from upbeat promotional tones and dramatic narration to empathetic, conversational support. Powered by our MultiNative technology, Murf voices can speak multiple languages, seamlessly switching mid-sentence while preserving native pronunciation patterns.
+
+Ideal for developers, Murf’s API makes it easy to build engaging audio experiences into any application.
+<Tip>
+ <code>MurfTTSService</code> supports real-time streaming and conversation context IDs.
+</Tip>
+
+<CardGroup cols={3}>
+  <Card
+    title="API Reference"
+    icon="code"
+    href="https://reference-server.pipecat.ai/en/latest/api/pipecat.services.murf.tts.html"
+  >
+    Complete API documentation and method details
+  </Card>
+  <Card
+    title="Murf Docs"
+    icon="book"
+    href="https://murf.ai/api/docs/api-reference/text-to-speech/stream-input"
+  >
+    Official Murf API documentation
+  </Card>
+  <Card
+    title="Example Code"
+    icon="play"
+    href="https://murf.ai/api/docs/text-to-speech/web-sockets#quickstart"
+  >
+    Working example with WebSocket quickstart
+  </Card>
+</CardGroup>
+
+## Installation
+
+To use Murf services, install the required dependencies:
+
+```bash
+pip install "pipecat-ai[murf]"
+```
+
+You'll also need to set up your Murf API key as an environment variable: `MURF_API_KEY`.
+
+<Tip>
+  Get your API key by signing up at [Murf.ai](https://murf.ai/api).
+</Tip>
+
+## Frames
+
+### Input
+
+- `TextFrame` - Text content to synthesize into speech
+- `TTSSpeakFrame` - Text that the TTS service should speak
+- `TTSUpdateSettingsFrame` - Runtime configuration updates
+- `StartInterruptionFrame` - Signals interruption request
+- `CancelFrame` - Signals cancellation request
+- `EndFrame` - Signals end of service
+
+### Output
+
+- `TTSStartedFrame` - Signals start of synthesis
+- `TTSAudioRawFrame` - Generated audio data chunks
+- `TTSStoppedFrame` - Signals completion of synthesis
+- `ErrorFrame` - Connection or processing errors
+
+## Features
+
+| Feature                | Support | Notes                                           |
+| --------------------- | ------- | ----------------------------------------------- |
+| **Streaming**         | ✅      | Real-time audio chunks via WebSocket            |
+| **Voice Styles**      | ✅      | Multiple styles including "Conversational"      |
+| **Rate Control**      | ✅      | Adjustable speech rate                          |
+| **Pitch Control**     | ✅      | Adjustable pitch                                |
+| **Pronunciation**     | ✅      | Custom pronunciation dictionary                  |
+| **Variation**         | ✅      | Control over pause, pitch, and speed variation  |
+| **Multi-language**    | ✅      | Support for multiple native locales             |
+| **Interruption**      | ✅      | Graceful handling via context management        |
+| **Metrics**           | ✅      | TTFB and usage metrics                          |
+
+## Usage Example
+
+Initialize the service with your API key and desired voice configuration:
+
+```python
+from pipecat.services.murf.tts import MurfTTSService
+
+# Configure WebSocket service
+tts = MurfTTSService(
+    api_key=os.getenv("MURF_API_KEY"),
+    params=MurfTTSService.InputParams(
+        voice_id="en-US-daniel",
+        style="Conversational",
+        rate=0,
+        pitch=0,
+        variation=1,
+        sample_rate=44100,
+        channel_type="MONO",
+        format="PCM"
+    )
+)
+
+# Use in pipeline
+pipeline = Pipeline([
+    transport.input(),
+    stt,
+    llm,
+    tts,
+    transport.output()
+])
+```
+
+### Input Parameters
+
+The `InputParams` class supports the following configuration options:
+
+| Parameter                   | Type    | Default         | Description                                    |
+| -------------------------- | ------- | --------------- | ---------------------------------------------- |
+| `voice_id`                 | str     | "en-US-daniel"  | Voice ID for synthesis                         |
+| `style`                    | str     | "Conversational"| Speech style                                   |
+| `rate`                     | int     | 0               | Speech rate adjustment                         |
+| `pitch`                    | int     | 0               | Speech pitch adjustment                        |
+| `variation`                | int     | 1               | Variation in pause, pitch, and speed           |
+| `pronunciation_dictionary` | dict    | None            | Custom word pronunciation mappings <br />Example: <br /> <code>\{ "live": \{ "type": "IPA", "pronunciation": "laɪv" }, "2024": \{ "type": "SAY_AS", "pronunciation": "twenty twenty-four" } }</code> |
+| `multi_native_locale`      | str     | None            | Specifies the language for the generated audio. Valid values: “en-US”, “en-UK”, “es-ES”, etc.                  |
+| `sample_rate`              | int     | 44100           | Audio sample rate, 8000, 16000, 24000, 32000, 44100, 48000                             |
+| `channel_type`             | str     | "MONO"          | Audio channel configuration, "MONO" or "STEREO"                  |
+| `format`                   | str     | "PCM"           | Audio output format, "PCM" or "MP3"                            |
+
+
+**Note**: Detailed documentation available at [Murf.ai](https://murf.ai/api/docs).
+
+### Dynamic Configuration
+
+You can update voice settings during runtime:
+
+```python
+from pipecat.frames.frames import TTSUpdateSettingsFrame
+
+await task.queue_frame(TTSUpdateSettingsFrame(
+    voice_id="different-voice",
+    style="Professional",
+    rate=1,
+    pitch=1
+))
+```
+
+## Metrics
+
+The service provides:
+
+- **Time to First Byte (TTFB)** - Latency measurement
+- **Processing Duration** - Synthesis time tracking
+- **Usage Metrics** - Character count and synthesis statistics
+
+<Info>
+  [Learn how to enable Metrics](/guides/features/metrics) in your Pipeline.
+</Info>