Azure · Howie Leung (howieleung) · Sep 15, 2026 · Sep 16, 2026 · Sep 17, 2026 · Sep 17, 2026
@@ -133,7 +133,7 @@ For each new sample, provide a one-line description of what it demonstrates. Rea
 Organize detected changes into these categories:
 
 ### Features Added
-- New sub-clients (e.g., "New `.beta.routines` sub-client with routine operations: `create_or_update`, `get`, `enable`, ...")
+- New sub-clients (e.g., "New `.beta.routines` sub-client for creating, scheduling, and managing routines.")
 - New methods on existing sub-clients (e.g., "New methods on `.beta.agents` for optimization jobs: `create_optimization_job`, `get_optimization_job`, ...")
 - New model classes that represent significant features (e.g., "Support integration of external Agents. See new `ExternalAgentDefinition` class.")
 - New properties on existing classes (e.g., "New optional `force` parameter on `agents.delete` method.")
@@ -201,7 +201,7 @@ Breaking changes in beta classes:
 
 **Guidelines for writing entries:**
 - For new methods: mention the sub-client and method name, briefly describe what it does. Only report the sync version — do not list both sync and async versions separately.
-- For new sub-clients: list all the methods it provides (sync versions only)
+- For new sub-clients: summarize the capability the sub-client provides; do not enumerate all of its methods.
 - For new tools: just mention the class name
 - For property changes: mention the class name and the affected property
 - For renames: show "X renamed to Y" format
@@ -282,9 +282,9 @@ Here's an example of a well-formatted changelog entry:
   * Optimization jobs: `create_optimization_job`, `get_optimization_job`, `list_optimization_jobs`, `cancel_optimization_job`, `list_optimization_candidates`.
   * Optimization candidate management: `list_optimization_candidates`, `get_optimization_candidate`, `get_optimization_candidate_config`, `get_optimization_candidate_results`, `get_candidate_file`, `promote_candidate`.
   * `stop_session` to stop a running agent session.
-* New `.beta.datasets` sub-client with data generation job operations: `create_generation_job`, `get_generation_job`, `list_generation_jobs`, `cancel_generation_job`, `delete_generation_job`.
-* New `.beta.models` sub-client to handle AI model weights: `create`, `list_versions`, `list`, `get`, `delete`, `update`, `pending_create_version`, `pending_upload`, `get_credentials`.
-* New `.beta.routines` sub-client with routine operations: `create_or_update`, `get`, `enable`, `disable`, `list`, `delete`, `list_runs`, `dispatch`.
+* New `.beta.datasets` sub-client for creating, monitoring, and managing data generation jobs.
+* New `.beta.models` sub-client for registering, versioning, and managing AI model weights.
+* New `.beta.routines` sub-client for creating, scheduling, and managing routines and their runs.
 * New methods on `.beta.evaluators` for evaluator generation jobs: `create_generation_job`, `get_generation_job`, `list_generation_jobs`, `cancel_generation_job`, `delete_generation_job`.
 * New methods on `.beta.memory_stores` to handle individual memory items: `create_memory`, `update_memory`, `list_memories`, `get_memory`, `delete_memory`.
 * New methods on `.beta.skills` for versioned skill management: `create`, `list_versions`, `get_version`, `download_version`, `delete_version`.

@@ -1,40 +1,49 @@
 # Release History
 
-## 2.6.1 (2026-09-14)
+## 2.7.0 (2026-09-18)
 
 ### Features Added
 
-* Added Voice Agents, a new agent kind for real-time, speech-to-speech conversational AI unified with the rest of the Agents API. Define a voice agent's model, audio, turn detection, greeting, and tools; manage it like any other agent; hold a live conversation with it over a WebSocket with barge-in and persisted conversation history/audio; and reach it through telephony (inbound bindings or outbound calls).
-* The core voice agent definition, as a new `kind="voice"` on `AgentDefinition`:
-  * Define a voice agent with `VoiceAgentDefinition`, configuring its model (`VoiceModelType`), audio input/output (`VoiceAgentAudioConfig`, `VoiceAgentAudioInputConfig`, `VoiceAgentAudioOutputConfig`), turn detection (`VoiceAgentTurnDetectionConfig` and its `VoiceAgentServerVadTurnDetection` / `VoiceAgentAzureSemanticVadTurnDetection` / `VoiceAgentAzureSemanticVadEnTurnDetection` / `VoiceAgentAzureSemanticVadMultilingualTurnDetection` variants), greeting (`VoiceAgentGreetingConfig` and its `VoiceAgentTemplateGreetingConfig` / `VoiceAgentLlmGeneratedGreetingConfig` variants), tools (`VoiceAgentTool`, `VoiceAgentFunctionTool`, `VoiceAgentMcpTool`, `VoiceAgentSystemTool` and its `VoiceAgentEndConversationSystemTool` variant, `VoiceAgentToolboxTool`), and avatar (`VoiceAgentAvatarConfig`). Manage it like any other agent through `project_client.agents` (`create_version`, `get`, `list`, `disable`/`enable`, `delete`).
-  * Added guided authoring via `project_client.beta.agents.create_from_prompt(GenerateVoiceAgentRequest(kind=AgentKind.VOICE, ...))`, which returns a service-generated starter definition that can be edited afterward through the standard `create_version`/`update` flow.
-  * Added a new `client.beta.voice_agents.realtime` / `async_client.beta.voice_agents.realtime` entry point for realtime speech-to-speech streaming. Use `with client.beta.voice_agents.realtime.connect(agent_name=...) as connection:` to open a WebSocket connection, `connection.send(...)` to send strongly-typed client events (or use the `connection.response`, `connection.conversation.item`, and `connection.session` helpers), and iterate over `connection` to receive strongly-typed server events (`RealtimeServerEvent*`). Conversation items exchanged with `connection.conversation.item.create(...)` are `RealtimeConversationItemMessageSystem`, `RealtimeConversationItemMessageUser`, `RealtimeConversationItemMessageAssistant`, `RealtimeConversationItemFunctionCall`, `RealtimeConversationItemFunctionCallOutput`, `RealtimeMCPApprovalResponse`, or a raw `Mapping[str, Any]`. The new types `BetaRealtime`, `BetaRealtimeConnection`, and `BetaRealtimeConnectionManager` (and their async equivalents `AsyncBetaRealtime`, `AsyncBetaRealtimeConnection`, `AsyncBetaRealtimeConnectionManager`) are exported from `azure.ai.projects.operations` / `azure.ai.projects.aio.operations`. These WebSocket clients identify themselves to the service the same way the generated HTTP surface does, via a standard Azure SDK `User-Agent` header and an `x-ms-client-sdk` query parameter for paths where the header isn't forwarded, so service telemetry can attribute this traffic to the SDK; a caller-supplied `User-Agent` in `extra_headers` still takes precedence. Requires the optional `websockets` package for the sync client, or `aiohttp` for the async client.
-  * Added the `.beta.voice_agents.conversations` operation group for reading back persisted voice-agent conversation transcripts and audio, for agents created with `store=True`.
-  * Added the underlying `RealtimeConversationItem*`, `RealtimeMCP*`, `RealtimeResponseUsage`, and related realtime event/session models used by the voice agent WebSocket protocol.
-* Telephony, WebRTC, and sub-agent consultation:
-  * Added telephony bindings so a voice agent can receive calls through Teams Phone or Twilio, through the new `.beta.voice_agents.telephony` operation group. `create_binding`/`get_binding`/`update_binding`/`delete_binding`/`list_bindings` manage the binding (`TelephonyBinding` and its `TeamsPhoneExtensionTelephonyBinding`/`TwilioTelephonyBinding` variants), and `list_calls`/`get_call`/`transfer_call`/`end_call`/`get_transfer_targets`/`replace_transfer_targets` manage in-progress and historical calls (`TelephonyCallRecord`, `TelephonyCallSummary`, `TelephonyCallTrace`, `TelephonyTransferTarget` and its `PSTNTelephonyTransferDestination`/`SipTelephonyTransferDestination`/`TeamsTelephonyTransferDestination` variants).
-  * Added outbound telephony call jobs on the same `.beta.voice_agents.telephony` operation group. `create_call_job`/`get_call_job`/`cancel_call_job` place and manage a single durable outbound call against a `TelephonyOutboundDestination` (`TelephonyCallJob`), originating directly from a `connection_name` (the Foundry connection to place the call from) and a `source` caller identity, with no pre-existing telephony binding required.
-  * Added an optional WebRTC transport for realtime voice sessions (`VoiceAgentTransport.WEBRTC`), where only SDP signaling travels over the WebSocket connection while media flows peer-to-peer. The new `VoiceAgentClientEventRtcCallSdpCreate`, `VoiceAgentServerEventRtcCallSdpCreated`, and `VoiceAgentServerEventRtcCallError` events carry the signaling exchange.
-  * Added the `.beta.voice_agents.conversations.get_generated_audio_item`/`download_generated_audio_item` methods for reading back a conversation item's *generated* audio, a subordinate artifact that can differ from what the listener heard when playback was interrupted, returning `VoiceGeneratedItemAudioResponse`.
-  * Added sub-agent consultation, letting a voice agent consult sibling Foundry text agents as background specialists mid-conversation, through the new `subagent_config` property on `VoiceAgentDefinition` (`VoiceAgentSubagentConfig`, `VoiceAgentSubagent`, `VoiceAgentSubagentResponsePolicy`), and the new `session.subagent.started`/`session.subagent.completed`/`session.subagent.aborted` realtime server events.
-  * Added an optional `conversation_engine` property on `VoiceAgentDefinition` (`VoiceConversationEngine`, `VoiceHostedAgentConversationEngine`) to delegate a voice agent's conversation handling to another hosted agent instead of configuring a model directly.
+* Added preview Voice Agent support, including voice agent definitions, realtime sessions and events, conversations, audio, and telephony models.
+* Added `.beta.agents.create_from_prompt()` to generate and create a Voice Agent from high-level inputs.
+* Added the `.beta.voice_agents.conversations` sub-client for managing voice conversations and retrieving their responses, conversation items, and audio.
+* Added `.beta.voice_agents.realtime.connect()` for sync and async realtime Voice Agent sessions with typed client and server events.
+* Added the `.beta.voice_agents.telephony` sub-client for managing calls, bindings, transfer targets, and durable outbound call jobs.
+* Added optional `harness` and `skills` properties to `PromptAgentDefinition`, with new GitHub Copilot harness, toolset, and skill-reference models.
+* Added invocation moderation through `RaiConfig.invocations_moderation` and `RaiInvocationModeration`.
+* Added `ToolboxesOperations.invoke_latest_toolbox_mcp()` and toolbox version metadata through `ToolboxObject.updated_at` and `ToolboxObject.versions`.
+* Added `DataGenerationJobOutputOptions.write_mode` for controlling dataset output writes and `TracesDataGenerationJobSource.trace_ids` for selecting explicit traces.
+* Made `TracesDataGenerationJobOptions.max_samples` optional.
+* Added read-only agent lifecycle properties `AgentDetails.configuration_state` and `AgentSessionResource.stopped_at`.
+
+### Breaking Changes
+
+Breaking changes in beta classes:
+
+* Removed the `max_samples` constructor argument and property from `DataGenerationJobOptions` and `SimulationSeedDataGenerationJobOptions`.
+* The `ToolboxObject` constructor now requires `updated_at` and `versions`.
 
 ### Dependency update
 
-* Added an optional dependency on `websockets` (sync `client.beta.voice_agents.realtime`) and `aiohttp` (async `async_client.beta.voice_agents.realtime`), required only when using the new voice agent realtime streaming APIs.
+* Added the optional `voice` dependency group, which installs `websockets` for sync realtime Voice Agent sessions and `aiohttp` for async sessions.
+
+### Sample updates
+
+* Added `sample_voice_agent_basic.py` under `samples/agents/voice/`, demonstrating the Voice Agent management lifecycle.
+* Added `sample_voice_agent_generate.py`, demonstrating guided Voice Agent authoring with `.beta.agents.create_from_prompt()`.
+* Added `sample_voice_agent_live_text_conversation.py`, demonstrating a persisted, typed realtime Voice Agent conversation.
+* Added `sample_voice_agent_live_audio_conversation_async.py`, demonstrating a hands-free realtime audio conversation with barge-in.
+* Added `sample_voice_agent_live_function_tool.py`, demonstrating client-side function execution during a realtime Voice Agent session.
+* Added `sample_voice_agent_read_conversation.py`, demonstrating how to read a persisted Voice Agent conversation and transcript.
+* Added `sample_voice_agent_read_conversation_audio.py`, demonstrating how to retrieve merged conversation audio and individual audio segments.
+* Added `sample_voice_agent_versions.py`, demonstrating Voice Agent version and draft management.
+* Added `sample_voice_agent_with_tools.py`, demonstrating audio configuration, tools, and self-deployed models.
+* Updated `sample_synthetic_multiturn_evaluation.py` to set the service-required simulation seed `max_samples` field through the model's mapping interface because `SimulationSeedDataGenerationJobOptions` no longer exposes it as a constructor argument.
+
+## 2.6.1 (2026-09-14)
 
 ### Sample updates
 
-* Added voice agent samples under `samples/agents/voice/`:
-  * `sample_voice_agent_basic.py` / `sample_voice_agent_basic_async.py` demonstrating the voice-agent management lifecycle: create, get, list, and delete.
-  * `sample_voice_agent_generate.py` demonstrating guided authoring of a voice agent via `.beta.agents.create_from_prompt` with `kind="voice"`.
-  * `sample_voice_agent_with_tools.py` demonstrating a richer voice agent definition: audio configuration, turn detection, greeting, and tools.
-  * `sample_voice_agent_versions.py` demonstrating voice-agent versioning: creating, drafting, listing, and publishing versions.
-  * `sample_voice_agent_live_text_conversation.py` / `sample_voice_agent_live_text_conversation_async.py` demonstrating a live, typed conversation with a voice agent over `client.beta.voice_agents.realtime`/`async_client.beta.voice_agents.realtime`.
-  * `sample_voice_agent_live_audio_conversation_async.py` demonstrating a hands-free, bidirectional live audio conversation over `async_client.beta.voice_agents.realtime`.
-  * `sample_voice_agent_live_function_tool.py` demonstrating handling a client-executed function tool during a live voice-agent session.
-  * `sample_voice_agent_read_conversation.py` demonstrating reading a persisted voice conversation's transcript back via `.beta.voice_agents.conversations`.
-  * `sample_voice_agent_read_conversation_audio.py` demonstrating reading a persisted voice conversation's audio, both the merged whole-call recording and a single turn's segment, via `.beta.voice_agents.conversations`.
 * Added `sample_agent_insights_on_demand.py` and `sample_agent_insights_scheduled.py` under `samples/agent_insights/` to demonstrate on-demand analysis, insight lifecycle updates, and scheduled monitor setup with temporary external agents, fictional traces, and owned-resource cleanup.
 * Added `sample_agent_web_iq.py` under `samples/agents/tools/`, demonstrating a Prompt Agent using the `WebIQPreviewTool`.