[VoiceLive] Fix MCP error (Azure#47325)

* [VoiceLive] Fix MCP error

* update code owner

* update spell words

* update format

---------

Co-authored-by: Xiting Zhang <[email protected]>

Author: xitzhang

.github/CODEOWNERS

@@ -92,7 +92,7 @@
 /sdk/ai/azure-ai-inference/                                        @dargilco @jhakulin @glharper @Azure/azure-java-sdk
 
 # PRLabel: %Voice Live
-/sdk/ai/azure-ai-voicelive/                                        @rhurey @xitzhang
+/sdk/ai/azure-ai-voicelive/                                        @rhurey @xitzhang @amber-yujueWang
 
 # ServiceLabel: %AKS
 # ServiceOwners:                                                   @Azure/aks-pm

.vscode/cspell.json

@@ -873,7 +873,9 @@
       "words": [
         "Dexec",
         "viseme",
-        "VISEME"
+        "VISEME",
+        "webrtc",
+        "WEBRTC"
       ]
     },
     {

sdk/ai/azure-ai-voicelive/CHANGELOG.md

@@ -4,11 +4,27 @@
 
 ### Features Added
 
-### Breaking Changes
-
-### Bugs Fixed
-
-### Other Changes
+- Added image input support for multimodal conversations:
+  - `RequestImageContentPart` for including images in conversation messages with URL references
+  - `RequestImageContentPartDetail` enum for controlling image detail level (auto, low, high)
+  - `ContentPartType.INPUT_IMAGE` discriminator for image content parts
+- Added avatar configuration enhancements:
+  - `AvatarConfiguration` class for configuring avatar streaming and behavior with ICE servers, character selection, style, and video parameters
+  - `AvatarConfigTypes` enum for video and photo avatar types
+  - `AvatarOutputProtocol` enum supporting WebRTC and WebSocket protocols
+  - `PhotoAvatarBaseModes` enum with VASA-1 model support
+- Added token usage tracking improvements:
+  - `CachedTokenDetails` for tracking cached text, audio, and image tokens
+  - Enhanced `InputTokenDetails` with image token tracking and cached token details
+- Added MCP call lifecycle events:
+  - `ServerEventResponseMcpCallInProgress` for tracking ongoing MCP calls
+  - `ServerEventResponseMcpCallCompleted` for successful MCP call completion
+  - `ServerEventResponseMcpCallFailed` for failed MCP calls
+- Added two new OpenAI voices: `OpenAIVoiceName.MARIN` and `OpenAIVoiceName.CEDAR`
+- Enhanced `AzurePersonalVoice` with additional customization options:
+  - Custom lexicon URL support for pronunciation customization
+  - Locale preferences with `preferLocales` for multilingual scenarios
+  - Voice style, pitch, rate, and volume controls for fine-tuned voice characteristics
 
 ## 1.0.0-beta.2 (2025-11-14)
 

sdk/ai/azure-ai-voicelive/src/main/java/com/azure/ai/voicelive/models/AvatarConfigTypes.java

@@ -0,0 +1,57 @@
+// Copyright (c) Microsoft Corporation. All rights reserved.
+// Licensed under the MIT License.
+// Code generated by Microsoft (R) TypeSpec Code Generator.
+package com.azure.ai.voicelive.models;
+
+import com.azure.core.annotation.Generated;
+import com.azure.core.util.ExpandableStringEnum;
+import java.util.Collection;
+
+/**
+ * Avatar config types.
+ */
+public final class AvatarConfigTypes extends ExpandableStringEnum<AvatarConfigTypes> {
+
+    /**
+     * Video avatar.
+     */
+    @Generated
+    public static final AvatarConfigTypes VIDEO_AVATAR = fromString("video-avatar");
+
+    /**
+     * Photo avatar.
+     */
+    @Generated
+    public static final AvatarConfigTypes PHOTO_AVATAR = fromString("photo-avatar");
+
+    /**
+     * Creates a new instance of AvatarConfigTypes value.
+     *
+     * @deprecated Use the {@link #fromString(String)} factory method.
+     */
+    @Generated
+    @Deprecated
+    public AvatarConfigTypes() {
+    }
+
+    /**
+     * Creates or finds a AvatarConfigTypes from its string representation.
+     *
+     * @param name a name to look for.
+     * @return the corresponding AvatarConfigTypes.
+     */
+    @Generated
+    public static AvatarConfigTypes fromString(String name) {
+        return fromString(name, AvatarConfigTypes.class);
+    }
+
+    /**
+     * Gets known AvatarConfigTypes values.
+     *
+     * @return known AvatarConfigTypes values.
+     */
+    @Generated
+    public static Collection<AvatarConfigTypes> values() {
+        return values(AvatarConfigTypes.class);
+    }
+}

sdk/ai/azure-ai-voicelive/src/main/java/com/azure/ai/voicelive/models/AvatarConfiguration.java

@@ -155,9 +155,13 @@ public JsonWriter toJson(JsonWriter jsonWriter) throws IOException {
         jsonWriter.writeStartObject();
         jsonWriter.writeStringField("character", this.character);
         jsonWriter.writeBooleanField("customized", this.customized);
+        jsonWriter.writeStringField("type", this.type == null ? null : this.type.toString());
         jsonWriter.writeArrayField("ice_servers", this.iceServers, (writer, element) -> writer.writeJson(element));
         jsonWriter.writeStringField("style", this.style);
+        jsonWriter.writeStringField("model", this.model == null ? null : this.model.toString());
         jsonWriter.writeJsonField("video", this.video);
+        jsonWriter.writeStringField("output_protocol",
+            this.outputProtocol == null ? null : this.outputProtocol.toString());
         return jsonWriter.writeEndObject();
     }
 
@@ -175,31 +179,127 @@ public static AvatarConfiguration fromJson(JsonReader jsonReader) throws IOExcep
         return jsonReader.readObject(reader -> {
             String character = null;
             boolean customized = false;
+            AvatarConfigTypes type = null;
             List<IceServer> iceServers = null;
             String style = null;
+            PhotoAvatarBaseModes model = null;
             VideoParams video = null;
+            AvatarOutputProtocol outputProtocol = null;
             while (reader.nextToken() != JsonToken.END_OBJECT) {
                 String fieldName = reader.getFieldName();
                 reader.nextToken();
                 if ("character".equals(fieldName)) {
                     character = reader.getString();
                 } else if ("customized".equals(fieldName)) {
                     customized = reader.getBoolean();
+                } else if ("type".equals(fieldName)) {
+                    type = AvatarConfigTypes.fromString(reader.getString());
                 } else if ("ice_servers".equals(fieldName)) {
                     iceServers = reader.readArray(reader1 -> IceServer.fromJson(reader1));
                 } else if ("style".equals(fieldName)) {
                     style = reader.getString();
+                } else if ("model".equals(fieldName)) {
+                    model = PhotoAvatarBaseModes.fromString(reader.getString());
                 } else if ("video".equals(fieldName)) {
                     video = VideoParams.fromJson(reader);
+                } else if ("output_protocol".equals(fieldName)) {
+                    outputProtocol = AvatarOutputProtocol.fromString(reader.getString());
                 } else {
                     reader.skipChildren();
                 }
             }
             AvatarConfiguration deserializedAvatarConfiguration = new AvatarConfiguration(character, customized);
+            deserializedAvatarConfiguration.type = type;
             deserializedAvatarConfiguration.iceServers = iceServers;
             deserializedAvatarConfiguration.style = style;
+            deserializedAvatarConfiguration.model = model;
             deserializedAvatarConfiguration.video = video;
+            deserializedAvatarConfiguration.outputProtocol = outputProtocol;
             return deserializedAvatarConfiguration;
         });
     }
+
+    /*
+     * Type of avatar to use.
+     */
+    @Generated
+    private AvatarConfigTypes type;
+
+    /*
+     * Base model to use for the avatar. Required for photo avatar.
+     */
+    @Generated
+    private PhotoAvatarBaseModes model;
+
+    /*
+     * Output protocol for avatar streaming. Default is 'webrtc'.
+     */
+    @Generated
+    private AvatarOutputProtocol outputProtocol;
+
+    /**
+     * Get the type property: Type of avatar to use.
+     *
+     * @return the type value.
+     */
+    @Generated
+    public AvatarConfigTypes getType() {
+        return this.type;
+    }
+
+    /**
+     * Set the type property: Type of avatar to use.
+     *
+     * @param type the type value to set.
+     * @return the AvatarConfiguration object itself.
+     */
+    @Generated
+    public AvatarConfiguration setType(AvatarConfigTypes type) {
+        this.type = type;
+        return this;
+    }
+
+    /**
+     * Get the model property: Base model to use for the avatar. Required for photo avatar.
+     *
+     * @return the model value.
+     */
+    @Generated
+    public PhotoAvatarBaseModes getModel() {
+        return this.model;
+    }
+
+    /**
+     * Set the model property: Base model to use for the avatar. Required for photo avatar.
+     *
+     * @param model the model value to set.
+     * @return the AvatarConfiguration object itself.
+     */
+    @Generated
+    public AvatarConfiguration setModel(PhotoAvatarBaseModes model) {
+        this.model = model;
+        return this;
+    }
+
+    /**
+     * Get the outputProtocol property: Output protocol for avatar streaming. Default is 'webrtc'.
+     *
+     * @return the outputProtocol value.
+     */
+    @Generated
+    public AvatarOutputProtocol getOutputProtocol() {
+        return this.outputProtocol;
+    }
+
+    /**
+     * Set the outputProtocol property: Output protocol for avatar streaming. Default is 'webrtc'.
+     *
+     * @param outputProtocol the outputProtocol value to set.
+     * @return the AvatarConfiguration object itself.
+     */
+    @Generated
+    public AvatarConfiguration setOutputProtocol(AvatarOutputProtocol outputProtocol) {
+        this.outputProtocol = outputProtocol;
+        return this;
+    }
 }

sdk/ai/azure-ai-voicelive/src/main/java/com/azure/ai/voicelive/models/AvatarOutputProtocol.java

@@ -0,0 +1,57 @@
+// Copyright (c) Microsoft Corporation. All rights reserved.
+// Licensed under the MIT License.
+// Code generated by Microsoft (R) TypeSpec Code Generator.
+package com.azure.ai.voicelive.models;
+
+import com.azure.core.annotation.Generated;
+import com.azure.core.util.ExpandableStringEnum;
+import java.util.Collection;
+
+/**
+ * Avatar config output protocols.
+ */
+public final class AvatarOutputProtocol extends ExpandableStringEnum<AvatarOutputProtocol> {
+
+    /**
+     * WebRTC protocol, output the audio/video streams via WebRTC.
+     */
+    @Generated
+    public static final AvatarOutputProtocol WEBRTC = fromString("webrtc");
+
+    /**
+     * WebSocket protocol, output the video frames over WebSocket.
+     */
+    @Generated
+    public static final AvatarOutputProtocol WEBSOCKET = fromString("websocket");
+
+    /**
+     * Creates a new instance of AvatarOutputProtocol value.
+     *
+     * @deprecated Use the {@link #fromString(String)} factory method.
+     */
+    @Generated
+    @Deprecated
+    public AvatarOutputProtocol() {
+    }
+
+    /**
+     * Creates or finds a AvatarOutputProtocol from its string representation.
+     *
+     * @param name a name to look for.
+     * @return the corresponding AvatarOutputProtocol.
+     */
+    @Generated
+    public static AvatarOutputProtocol fromString(String name) {
+        return fromString(name, AvatarOutputProtocol.class);
+    }
+
+    /**
+     * Gets known AvatarOutputProtocol values.
+     *
+     * @return known AvatarOutputProtocol values.
+     */
+    @Generated
+    public static Collection<AvatarOutputProtocol> values() {
+        return values(AvatarOutputProtocol.class);
+    }
+}