docs: multilingual guides

Author: goosewin

fern/assistants/examples/multilingual-agent.mdx

@@ -15,9 +15,9 @@ Configure your voice assistant to communicate in multiple languages with automat
 - Design language-aware system prompts
 - Test and optimize multilingual performance
 
-<Warning>
-**Provider Limitations:** Most transcription providers don't support true multilingual mode. Only **Google STT** currently supports automatic language detection and switching within a single conversation.
-</Warning>
+<Note>
+**Multilingual Support:** Multiple providers support automatic language detection. **Deepgram** (Nova 2, Nova 3 with "Multi" setting) and **Google STT** (with "Multilingual" setting) both offer automatic language detection for seamless multilingual conversations.
+</Note>
 
 ## Configure automatic language detection
 
@@ -28,12 +28,10 @@ Set up your transcriber to automatically detect and process multiple languages.
     1. Navigate to **Assistants** in your [Vapi Dashboard](https://dashboard.vapi.ai/)
     2. Create a new assistant or edit an existing one
     3. In the **Transcriber** section:
-       - **Provider**: Select `Google` (only provider with true multilingual support)
-       - **Model**: Choose Flash or Flash Lite (recommended for speed)
-       - **Language Detection**: Set to `Automatic multiple languages`
-    4. **Alternative for limited multilingual**: 
-       - **Deepgram**: Only supports English and Spanish in "Multi" mode
-       - **Other providers**: Single language only, no auto-detection
+       - **Provider**: Select `Deepgram` (recommended) or `Google`
+       - **Model**: For Deepgram, choose `Nova 2` or `Nova 3`; for Google, choose `Latest`
+       - **Language**: Set to `Multi` (Deepgram) or `Multilingual` (Google)
+    4. **Other providers**: Single language only, no automatic detection
     5. Click **Save** to apply the configuration
   </Tab>
   <Tab title="TypeScript (Server SDK)">
@@ -42,21 +40,21 @@ Set up your transcriber to automatically detect and process multiple languages.
 
     const vapi = new VapiClient({ token: "YOUR_VAPI_API_KEY" });
 
-    // Recommended: Google for true multilingual support
+    // Recommended: Deepgram for multilingual support
     const assistant = await vapi.assistants.create({
       name: "Multilingual Assistant",
       transcriber: {
-        provider: "google",
-        model: "gemini-2.0-flash-lite", // or "gemini-2.0-flash" for better accuracy
-        languageBehaviour: "automatic multiple languages"
+        provider: "deepgram",
+        model: "nova-2", // or "nova-3"
+        language: "multi"
       }
     });
 
-    // Alternative: Deepgram for English + Spanish only
-    const limitedMultilingual = {
-      provider: "deepgram",
-      model: "nova-2",
-      languageBehaviour: "Multi" // Only en + es supported
+    // Alternative: Google for multilingual support
+    const googleMultilingual = {
+      provider: "google",
+      model: "latest",
+      language: "multilingual"
     };
     ```
   </Tab>
@@ -67,56 +65,56 @@ Set up your transcriber to automatically detect and process multiple languages.
 
     client = Vapi(token=os.getenv("VAPI_API_KEY"))
 
-    # Recommended: Google for true multilingual support
+    # Recommended: Deepgram for multilingual support
     assistant = client.assistants.create(
         name="Multilingual Assistant",
         transcriber={
-            "provider": "google",
-            "model": "gemini-2.0-flash-lite",  # or "gemini-2.0-flash" for better accuracy
-            "languageBehaviour": "automatic multiple languages"
+            "provider": "deepgram",
+            "model": "nova-2",  # or "nova-3"
+            "language": "multi"
         }
     )
 
-    # Alternative: Deepgram for English + Spanish only
-    limited_multilingual = {
-        "provider": "deepgram",
-        "model": "nova-2",
-        "languageBehaviour": "Multi"  # Only en + es supported
+    # Alternative: Google for multilingual support
+    google_multilingual = {
+        "provider": "google",
+        "model": "latest",
+        "language": "multilingual"
     }
     ```
   </Tab>
   <Tab title="cURL">
     ```bash
-    # Recommended: Google for true multilingual support
+    # Recommended: Deepgram for multilingual support
     curl -X POST "https://api.vapi.ai/assistant" \
          -H "Authorization: Bearer $VAPI_API_KEY" \
          -H "Content-Type: application/json" \
          -d '{
            "name": "Multilingual Assistant",
            "transcriber": {
-             "provider": "google",
-             "model": "gemini-2.0-flash-lite",
-             "languageBehaviour": "automatic multiple languages"
+             "provider": "deepgram",
+             "model": "nova-2",
+             "language": "multi"
            }
          }'
 
-    # Alternative: Deepgram for English + Spanish only
+    # Alternative: Google for multilingual support
     curl -X POST "https://api.vapi.ai/assistant" \
          -H "Authorization: Bearer $VAPI_API_KEY" \
          -H "Content-Type: application/json" \
          -d '{
            "transcriber": {
-             "provider": "deepgram",
-             "model": "nova-2",
-             "languageBehaviour": "Multi"
+             "provider": "google",
+             "model": "latest",
+             "language": "multilingual"
            }
          }'
     ```
   </Tab>
 </Tabs>
 
 <Note>
-**Google STT Performance:** While Google is slower than other providers, use the `gemini-2.0-flash-lite` or `gemini-2.0-flash` models for better speed. These lighter models provide the best balance of multilingual support and performance.
+**Provider Performance:** **Deepgram** offers the best balance of speed and multilingual accuracy. **Google** provides broader language support but may be slower. Both providers support automatic language detection within conversations.
 </Note>
 
 ## Set up multilingual voices
@@ -460,8 +458,8 @@ Validate your configuration with different languages and scenarios.
 
 | Provider | Multilingual Support | Languages | Notes |
 |----------|---------------------|-----------|-------|
-| **Google STT** | ✅ Full auto-detection | 125+ | **Only provider with true multilingual mode**. Use Flash or Flash Lite models for speed. |
-| **Deepgram** | ⚠️ Limited | English + Spanish only | "Multi" mode supports only 2 languages |
+| **Deepgram** | ✅ Full auto-detection | 100+ | **Recommended**: Nova 2/Nova 3 with "Multi" language setting |
+| **Google STT** | ✅ Full auto-detection | 125+ | Latest models with "Multilingual" language setting |
 | **Assembly AI** | ❌ English only | English | No multilingual support |
 | **Azure STT** | ❌ Single language | 100+ | Many languages, but no auto-detection |
 | **OpenAI Whisper** | ❌ Single language | 90+ | Many languages, but no auto-detection |
@@ -483,10 +481,10 @@ Validate your configuration with different languages and scenarios.
 <AccordionGroup>
   <Accordion title="Language detection is inaccurate">
     **Solutions:**
-    - Use Google STT (only provider with reliable multilingual support)
-    - For Deepgram, stick to English + Spanish combinations only
-    - Use higher-quality audio preprocessing
+    - Use Deepgram (Nova 2/Nova 3 with "Multi") or Google STT (with "Multilingual")
+    - Ensure high-quality audio input for better detection accuracy
     - Test with native speakers of target languages
+    - Consider provider-specific language combinations for optimal results
   </Accordion>
 
   <Accordion title="Assistant doesn't realize it can speak multiple languages">
@@ -497,12 +495,12 @@ Validate your configuration with different languages and scenarios.
     - Avoid generic "multilingual" statements without specifics
   </Accordion>
 
-  <Accordion title="Google STT is too slow">
+  <Accordion title="Transcription is too slow">
     **Solutions:**
-    - Use `gemini-2.0-flash-lite` or `gemini-2.0-flash` models instead of standard models
-    - Consider the speed vs accuracy tradeoff
-    - For English + Spanish only, use Deepgram's "Multi" mode
-    - Optimize your audio quality to improve processing speed
+    - Use Deepgram Nova 2/Nova 3 for optimal speed and multilingual support
+    - For Google STT, use latest models for better performance
+    - Consider the speed vs accuracy tradeoff for your use case
+    - Optimize audio quality and format to improve processing speed
   </Accordion>
 
   <Accordion title="Voice quality varies between languages">

fern/customization/multilingual.mdx

@@ -258,6 +258,9 @@ navigation:
               - page: Order management
                 path: workflows/examples/ecommerce-order-management.mdx
                 icon: fa-light fa-shopping-cart
+              - page: Multilingual support
+                path: workflows/examples/multilingual-support.mdx
+                icon: fa-light fa-globe
 
       - section: Best practices
         contents:

fern/docs.yml

@@ -41,6 +41,22 @@ slug: guides
     <br />
     Create an outbound sales agent that can schedule appointments automatically
   </Card>
+  <Card title="Multilingual Support Workflow" icon="globe" href="/workflows/examples/multilingual-support">
+    <div className='absolute top-4 right-4'>
+      <Icon icon="arrow-up-right-from-square" />
+    </div>
+    <div class="vapi-badge vapi-badge-workflow">Built with Workflows</div>
+    <br />
+    Build a structured multilingual support workflow with language selection and dedicated conversation paths
+  </Card>
+  <Card title="Dynamic Multilingual Agent" icon="language" href="/assistants/examples/multilingual-agent">
+    <div className='absolute top-4 right-4'>
+      <Icon icon="arrow-up-right-from-square" />
+    </div>
+    <div class="vapi-badge vapi-badge-assistant">Built with Assistants</div>
+    <br />
+    Build a dynamic agent with automatic language detection and real-time language switching
+  </Card>
   <Card title="Docs Agent" icon="book-open" href="/assistants/examples/docs-agent">
     <div className='absolute top-4 right-4'>
       <Icon icon="arrow-up-right-from-square" />

fern/guides.mdx

@@ -7,7 +7,7 @@ slug: providers/model/openai
 
 **What is OpenAI?**
 
-OpenAI is a leading artificial intelligence research and deployment company dedicated to ensuring that artificial general intelligence (AGI) benefits all of humanity. Founded with the mission to create safe and highly capable AI systems, OpenAI has made significant strides in AI research, producing groundbreaking models like GPT-4, DALL-E, and Codex. These innovations have not only advanced the field of AI but also transformed various industries by providing powerful tools for natural language processing, image generation, and programming assistance.
+OpenAI is a leading artificial intelligence research and deployment company dedicated to ensuring that artificial general intelligence (AGI) benefits all of humanity. Founded with the mission to create safe and highly capable AI systems, OpenAI has made significant strides in AI research, producing groundbreaking models like GPT-4o, DALL-E, and Codex. These innovations have not only advanced the field of AI but also transformed various industries by providing powerful tools for natural language processing, image generation, and programming assistance.
 
 **The Evolution of AI Research:**
 
@@ -19,7 +19,7 @@ OpenAI offers a range of AI-driven products and services designed to meet divers
 
 **GPT Models:**
 
-- OpenAI’s Generative Pre-trained Transformer (GPT) models, including the latest GPT-4, are state-of-the-art in natural language processing. These models can generate human-like text, answer questions, summarize information, and perform various language tasks with high accuracy. GPT-4, in particular, represents a significant leap in AI capabilities, offering improved coherence, context understanding, and creativity.
+- OpenAI’s Generative Pre-trained Transformer (GPT) models, including the latest GPT-4o, are state-of-the-art in natural language processing. These models can generate human-like text, answer questions, summarize information, and perform various language tasks with high accuracy. GPT-4o, in particular, represents a significant leap in AI capabilities, offering improved coherence, context understanding, and creativity.
 
 **DALL-E:**
 

fern/providers/model/openai.mdx

@@ -7,7 +7,7 @@ slug: providers/voice/openai
 
 **What is OpenAI?**
 
-OpenAI is a leading artificial intelligence research and deployment company dedicated to ensuring that artificial general intelligence (AGI) benefits all of humanity. Founded with the mission to create safe and highly capable AI systems, OpenAI has made significant strides in AI research, producing groundbreaking models like GPT-4, DALL-E, and Codex. These innovations have not only advanced the field of AI but also transformed various industries by providing powerful tools for natural language processing, image generation, and programming assistance.
+OpenAI is a leading artificial intelligence research and deployment company dedicated to ensuring that artificial general intelligence (AGI) benefits all of humanity. Founded with the mission to create safe and highly capable AI systems, OpenAI has made significant strides in AI research, producing groundbreaking models like GPT-4o, DALL-E, and Codex. These innovations have not only advanced the field of AI but also transformed various industries by providing powerful tools for natural language processing, image generation, and programming assistance.
 
 **The Evolution of AI Research:**
 
@@ -19,7 +19,7 @@ OpenAI offers a range of AI-driven products and services designed to meet divers
 
 **GPT Models:**
 
-- OpenAI’s Generative Pre-trained Transformer (GPT) models, including the latest GPT-4, are state-of-the-art in natural language processing. These models can generate human-like text, answer questions, summarize information, and perform various language tasks with high accuracy. GPT-4, in particular, represents a significant leap in AI capabilities, offering improved coherence, context understanding, and creativity.
+- OpenAI’s Generative Pre-trained Transformer (GPT) models, including the latest GPT-4o, are state-of-the-art in natural language processing. These models can generate human-like text, answer questions, summarize information, and perform various language tasks with high accuracy. GPT-4o, in particular, represents a significant leap in AI capabilities, offering improved coherence, context understanding, and creativity.
 
 **DALL-E:**
 

fern/providers/voice/openai.mdx

@@ -0,0 +1,16 @@
+customer_id,name,email,phone,preferred_language,region,account_status,support_tier
+C001,Sarah Johnson,[email protected],+1-555-0101,english,north_america,active,premium
+C002,María González,[email protected],+34-600-123456,spanish,spain,active,standard
+C003,Jean Dubois,[email protected],+33-1-23456789,french,france,active,premium
+C004,Carlos Restrepo,[email protected],+52-55-12345678,spanish,mexico,active,standard
+C005,Emily Chen,[email protected],+1-555-0102,english,north_america,active,enterprise
+C006,Sophie Martin,[email protected],+1-514-1234567,french,canada,active,premium
+C007,Antonio Silva,[email protected],+34-91-2345678,spanish,spain,inactive,standard
+C008,Michael Thompson,[email protected],+44-20-12345678,english,uk,active,standard
+C009,Luisa Fernández,[email protected],+57-1-3456789,spanish,colombia,active,premium
+C010,Pierre Leblanc,[email protected],+33-4-56789012,french,france,active,standard
+C011,Jennifer Davis,[email protected],+1-555-0103,english,north_america,active,standard
+C012,Isabel Rodríguez,[email protected],+52-33-45678901,spanish,mexico,suspended,standard
+C013,François Moreau,[email protected],+1-418-5678901,french,canada,active,enterprise
+C014,David Wilson,[email protected],+1-555-0104,english,north_america,active,premium
+C015,Carmen López,[email protected],+34-93-6789012,spanish,spain,active,standard 

fern/static/images/workflows/examples/multilingual-support.png

@@ -0,0 +1,11 @@
+product_id,name_en,name_es,name_fr,category,price_usd,description_en,description_es,description_fr,availability
+P001,CloudSync Pro,CloudSync Pro,CloudSync Pro,cloud_storage,29.99,"Professional cloud storage with 1TB space and advanced sharing","Almacenamiento en la nube profesional con 1TB de espacio y compartición avanzada","Stockage cloud professionnel avec 1 To d'espace et partage avancé",available
+P002,SecureVPN Enterprise,VPN Segura Empresarial,VPN Sécurisé Entreprise,security,49.99,"Enterprise-grade VPN with global servers and 24/7 support","VPN de nivel empresarial con servidores globales y soporte 24/7","VPN de niveau entreprise avec serveurs mondiaux et support 24h/24",available
+P003,DataAnalytics Suite,Suite de Análisis de Datos,Suite d'Analyse de Données,analytics,99.99,"Complete data analytics platform with AI-powered insights","Plataforma completa de análisis de datos con insights impulsados por IA","Plateforme d'analyse de données complète avec des insights alimentés par l'IA",available
+P004,Mobile App Builder,Constructor de Apps Móviles,Constructeur d'Apps Mobiles,development,79.99,"No-code mobile app development platform","Plataforma de desarrollo de aplicaciones móviles sin código","Plateforme de développement d'applications mobiles sans code",available
+P005,Smart Backup Pro,Backup Inteligente Pro,Sauvegarde Intelligente Pro,backup,19.99,"Automated backup solution with intelligent scheduling","Solución de respaldo automatizada con programación inteligente","Solution de sauvegarde automatisée avec planification intelligente",available
+P006,Team Collaboration Hub,Centro de Colaboración,Hub de Collaboration,collaboration,39.99,"All-in-one team collaboration platform","Plataforma de colaboración en equipo todo-en-uno","Plateforme de collaboration d'équipe tout-en-un",available
+P007,AI Assistant Pro,Asistente IA Pro,Assistant IA Pro,ai,149.99,"Advanced AI assistant for business automation","Asistente de IA avanzado para automatización empresarial","Assistant IA avancé pour l'automatisation des entreprises",beta
+P008,Network Monitor,Monitor de Red,Moniteur Réseau,monitoring,59.99,"Real-time network monitoring and alerting","Monitoreo de red en tiempo real y alertas","Surveillance réseau en temps réel et alertes",available
+P009,Database Optimizer,Optimizador de Base de Datos,Optimiseur de Base de Données,database,89.99,"Database performance optimization tool","Herramienta de optimización de rendimiento de base de datos","Outil d'optimisation des performances de base de données",available
+P010,Code Review Assistant,Asistente de Revisión de Código,Assistant de Révision de Code,development,69.99,"Automated code review with quality insights","Revisión automatizada de código con insights de calidad","Révision automatisée du code avec des insights qualité",coming_soon