ref

Author: Guikingone

README.md

@@ -13,7 +13,6 @@ Symfony AI consists of several lower and higher level **components** and the res
   * **[Agent](src/agent/README.md)**: Framework for building AI agents that can interact with users and perform tasks.
   * **[Chat](src/chat/README.md)**: An unified interface to send messages to agents and store long-term context.
   * **[Store](src/store/README.md)**: Data storage abstraction with indexing and retrieval for AI applications.
-  * **[Voice](src/voice/README.md)**: An unified interface to provide TTS / STT / STS for agents and more.
 * **Bundles**
   * **[AI Bundle](src/ai-bundle/README.md)**: Symfony integration for AI Platform, Store and Agent components.
   * **[MCP Bundle](src/mcp-bundle/README.md)**: Symfony integration for official MCP SDK, allowing them to act as MCP servers or clients.

docs/components/platform.rst

@@ -501,6 +501,9 @@ This allows fast and isolated testing of AI-powered features without relying on
 
     This requires `cURL` and the `ext-curl` extension to be installed.
 
+Adding Voice
+~~~~~~~~~~~~
+
 Code Examples
 ~~~~~~~~~~~~~
 

docs/components/voice.rst

@@ -33,4 +33,4 @@
     'voice' => 'Dslrhjl3ZpzrctukrQSN', // Brad (https://elevenlabs.io/app/voice-library?voiceId=Dslrhjl3ZpzrctukrQSN)
 ]);
 
-echo $answer->asVoice();
+echo $result->asVoice();

examples/voice/agent-eleven-labs-voice.php

@@ -25,7 +25,6 @@
         "phpstan/phpdoc-parser": "^2.1",
         "psr/log": "^3.0",
         "symfony/ai-platform": "@dev",
-        "symfony/ai-voice": "@dev",
         "symfony/clock": "^7.3|^8.0",
         "symfony/http-client": "^7.3|^8.0",
         "symfony/property-access": "^7.3|^8.0",

src/agent/composer.json

@@ -17,7 +17,7 @@
 use Symfony\AI\Platform\Message\MessageBag;
 use Symfony\AI\Platform\PlatformInterface;
 use Symfony\AI\Platform\Result\ResultInterface;
-use Symfony\AI\Voice\VoiceProviderInterface;
+use Symfony\AI\Platform\Voice\SpeechProviderInterface;
 
 /**
  * @author Christopher Hertel <[email protected]>
@@ -37,7 +37,7 @@ final class Agent implements AgentInterface
     /**
      * @var VoiceProviderInterface[]
      */
-    private array $voiceProviders;
+    private readonly array $voiceProviders;
 
     /**
      * @param InputProcessorInterface[]  $inputProcessors
@@ -88,7 +88,7 @@ public function call(MessageBag $messages, array $options = []): ResultInterface
 
         $output = new Output($model, $result, $messages, $options);
         array_map(static fn (OutputProcessorInterface $processor) => $processor->processOutput($output), $this->outputProcessors);
-        array_map(static fn (VoiceProviderInterface $provider) => $provider->addVoice($output), $this->voiceProviders);
+        array_map(static fn (SpeechProviderInterface $provider) => $provider->addVoice($output), $this->voiceProviders);
 
         return $output->getResult();
     }
@@ -122,8 +122,8 @@ private function initializeProcessors(iterable $processors, string $interface):
     private function initializeVoiceProviders(iterable $providers): array
     {
         foreach ($providers as $provider) {
-            if (!$provider instanceof VoiceProviderInterface) {
-                throw new InvalidArgumentException(\sprintf('Voice provider "%s" must implement "%s".', $provider::class, VoiceProviderInterface::class));
+            if (!$provider instanceof SpeechProviderInterface) {
+                throw new InvalidArgumentException(\sprintf('Speech provider "%s" must implement "%s".', $provider::class, SpeechProviderInterface::class));
             }
 
             if ($provider instanceof AgentAwareInterface) {

src/agent/src/Agent.php

@@ -13,6 +13,7 @@
 
 use Symfony\AI\Platform\Message\MessageBag;
 use Symfony\AI\Platform\Result\ResultInterface;
+use Symfony\AI\Platform\Speech\Speech;
 
 /**
  * @author Christopher Hertel <[email protected]>
@@ -27,7 +28,7 @@ public function __construct(
         private ResultInterface $result,
         private readonly MessageBag $messageBag,
         private readonly array $options = [],
-        private ?Voice $voice = null,
+        private ?Speech $speech = null,
     ) {
     }
 
@@ -59,13 +60,13 @@ public function getOptions(): array
         return $this->options;
     }
 
-    public function setVoice(?Voice $voice): void
+    public function setSpeech(?Speech $speech): void
     {
-        $this->voice = $voice;
+        $this->speech = $speech;
     }
 
-    public function getVoice(): ?Voice
+    public function getSpeech(): ?Speech
     {
-        return $this->voice;
+        return $this->speech;
     }
 }

src/agent/src/Output.php

@@ -102,7 +102,7 @@
 use Symfony\AI\Store\Indexer;
 use Symfony\AI\Store\IndexerInterface;
 use Symfony\AI\Store\StoreInterface;
-use Symfony\AI\Voice\Bridge\ElevenLabs\VoiceProvider as ElevenLabsVoiceProvider;
+use Symfony\AI\Voice\Bridge\ElevenLabs\ElevenLabsSpeechProvider as ElevenLabsVoiceProvider;
 use Symfony\AI\Voice\VoiceProviderInterface;
 use Symfony\Component\Clock\ClockInterface;
 use Symfony\Component\Config\Definition\Configurator\DefinitionConfigurator;
@@ -1774,7 +1774,7 @@ private function processVoiceProviderConfig(string $name, array $providers, Cont
     {
         if ('eleven_labs' === $name) {
             foreach ($providers as $type => $config) {
-                $definition = new Definition(ElevenLabsVoiceProvider::class);
+                $definition = new Definition(ElevenLabsSpeechProvider::class);
                 $definition
                     ->setLazy(true)
                     ->setArguments([

src/ai-bundle/src/AiBundle.php

@@ -0,0 +1,57 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <[email protected]>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\AI\Voice\Bridge\ElevenLabs;
+
+use Symfony\AI\Agent\Output;
+use Symfony\AI\Platform\Speech\Speech;
+use Symfony\AI\Platform\Speech\SpeechListenerInterface;
+use Symfony\AI\Platform\Speech\SpeechProviderInterface;
+use Symfony\AI\Platform\Platform;
+
+/**
+ * @author Guillaume Loulier <[email protected]>
+ */
+final class ElevenLabsSpeechProvider implements SpeechProviderInterface, SpeechListenerInterface
+{
+    public const ELEVEN_LABS_STT_MODEL = 'eleven_labs.stt_model';
+
+    public function __construct(
+        private readonly Platform $platform,
+        private readonly string $model,
+    ) {
+    }
+
+    public function addSpeech(Output $output): void
+    {
+        $result = $output->getResult();
+
+        $voice = $this->platform->invoke($this->model, $result->getContent());
+
+        $output->setSpeech(new Speech($result->getContent(), $voice));
+    }
+
+    public function listen(object|array|string $input, array $options): Speech
+    {
+        $model = $options[self::ELEVEN_LABS_STT_MODEL];
+
+        unset($options[self::ELEVEN_LABS_STT_MODEL]);
+
+        $text = $this->platform->invoke($model, $input);
+
+        return new Speech($input, $text);
+    }
+
+    public function supportListening(object|array|string $input, array $options): bool
+    {
+        return \array_key_exists(self::ELEVEN_LABS_STT_MODEL, $options);
+    }
+}

src/platform/src/Bridge/ElevenLabs/ElevenLabsSpeechProvider.php

@@ -18,6 +18,7 @@
 use Symfony\AI\Platform\ModelCatalog\ModelCatalogInterface;
 use Symfony\AI\Platform\Result\DeferredResult;
 use Symfony\AI\Platform\Result\RawResultInterface;
+use Symfony\AI\Platform\Speech\SpeechListenerInterface;
 
 /**
  * @author Christopher Hertel <[email protected]>
@@ -34,20 +35,28 @@ final class Platform implements PlatformInterface
      */
     private readonly array $resultConverters;
 
+    /**
+     * @var SpeechListenerInterface[]
+     */
+    private readonly iterable $speechListeners;
+
     /**
      * @param iterable<ModelClientInterface>     $modelClients
      * @param iterable<ResultConverterInterface> $resultConverters
+     * @param iterable<SpeechListenerInterface>  $speechListeners
      */
     public function __construct(
         iterable $modelClients,
         iterable $resultConverters,
+        iterable $speechListeners,
         private readonly ModelCatalogInterface $modelCatalog,
         private ?Contract $contract = null,
         private readonly ?EventDispatcherInterface $eventDispatcher = null,
     ) {
         $this->contract = $contract ?? Contract::create();
         $this->modelClients = $modelClients instanceof \Traversable ? iterator_to_array($modelClients) : $modelClients;
         $this->resultConverters = $resultConverters instanceof \Traversable ? iterator_to_array($resultConverters) : $resultConverters;
+        $this->speechListeners = $speechListeners instanceof \Traversable ? iterator_to_array($speechListeners) : $speechListeners;
     }
 
     public function invoke(string $model, array|string|object $input, array $options = []): DeferredResult
@@ -65,8 +74,9 @@ public function invoke(string $model, array|string|object $input, array $options
         }
 
         $result = $this->convertResult($model, $this->doInvoke($model, $payload, $options), $options);
+        $finalResult = $this->addSpeech($result, $payload, $options);
 
-        $event = new ResultEvent($model, $result, $options);
+        $event = new ResultEvent($model, $finalResult, $options);
         $this->eventDispatcher?->dispatch($event);
 
         return $event->getDeferredResult();
@@ -105,4 +115,21 @@ private function convertResult(Model $model, RawResultInterface $result, array $
 
         throw new RuntimeException(\sprintf('No ResultConverter registered for model "%s" with given input.', $model::class));
     }
+
+    private function addSpeech(DeferredResult $result, array|string $payload, array $options): DeferredResult
+    {
+        if ([] === $this->speechListeners) {
+            return $result;
+        }
+
+        foreach ($this->speechListeners as $speechListener) {
+            if ($speechListener->supportListening($payload, $options)) {
+                $result->setSpeech($speechListener->listen($payload, $options));
+
+                return $result;
+            }
+        }
+
+        throw new RuntimeException('No SpeechListener registered.');
+    }
 }