feat(realtime): add GA support for Realtime (#402)

Author: dkundel-openai

.changeset/sixty-hotels-care.md

@@ -0,0 +1,5 @@
+---
+'@openai/agents-realtime': patch
+---
+
+Add backgroundResult as an option to return tool results without triggering a new response

.changeset/tiny-maps-begin.md

@@ -1,6 +1,6 @@
 ---
-"@openai/agents-core": patch
-"@openai/agents-openai": patch
+'@openai/agents-core': patch
+'@openai/agents-openai': patch
 ---
 
 Fix #374 add connector support

.changeset/two-spiders-end.md

@@ -0,0 +1,7 @@
+---
+'@openai/agents-extensions': minor
+'@openai/agents-realtime': minor
+'@openai/agents-core': minor
+---
+
+moving realtime to the new GA API and add MCP support

AGENTS.md

@@ -57,6 +57,7 @@ Before submitting changes, ensure all checks pass:
   ```bash
   pnpm -r build-check
   ```
+  NEVER USE `-w` or other watch modes.
 - Run the full test suite:
   ```bash
   CI=1 pnpm test

docs/src/content/docs/guides/voice-agents/quickstart.mdx

@@ -47,11 +47,12 @@ import thinClientExample from '../../../../../../examples/docs/voice-agents/thin
    As this application will run in the user's browser, we need a secure way to connect to the model through the Realtime API. For this we can use an [ephemeral client key](https://platform.openai.com/docs/guides/realtime#creating-an-ephemeral-token) that should be generated on your backend server. For testing purposes you can also generate a key using `curl` and your regular OpenAI API key.
 
    ```bash
-   curl -X POST https://api.openai.com/v1/realtime/sessions \
+   curl -X POST https://api.openai.com/v1/realtime/client_secrets \
       -H "Authorization: Bearer $OPENAI_API_KEY" \
       -H "Content-Type: application/json" \
       -d '{
-        "model": "gpt-4o-realtime-preview-2025-06-03"
+        "type": "realtime",
+        "model": "gpt-realtime"
       }'
    ```
 
@@ -78,7 +79,7 @@ import thinClientExample from '../../../../../../examples/docs/voice-agents/thin
    import { RealtimeSession } from '@openai/agents-realtime';
 
    const session = new RealtimeSession(agent, {
-     model: 'gpt-4o-realtime-preview-2025-06-03',
+     model: 'gpt-realtime',
    });
    ```
 

docs/src/content/docs/ja/guides/voice-agents/quickstart.mdx

@@ -47,11 +47,12 @@ import thinClientExample from '../../../../../../../examples/docs/voice-agents/t
    このアプリケーションは ユーザー のブラウザで実行されるため、Realtime API を通じてモデルに安全に接続する必要があります。そのために、バックエンド サーバー で生成する [ephemeral client key](https://platform.openai.com/docs/guides/realtime#creating-an-ephemeral-token) を使用できます。テスト目的では、`curl` と通常の OpenAI API キーを使ってキーを生成することもできます。
 
    ```bash
-   curl -X POST https://api.openai.com/v1/realtime/sessions \
+   curl -X POST https://api.openai.com/v1/realtime/client_secrets \
       -H "Authorization: Bearer $OPENAI_API_KEY" \
       -H "Content-Type: application/json" \
       -d '{
-        "model": "gpt-4o-realtime-preview-2025-06-03"
+        "type": "realtime",
+        "model": "gpt-realtime"
       }'
    ```
 
@@ -78,7 +79,7 @@ import thinClientExample from '../../../../../../../examples/docs/voice-agents/t
    import { RealtimeSession } from '@openai/agents-realtime';
 
    const session = new RealtimeSession(agent, {
-     model: 'gpt-4o-realtime-preview-2025-06-03',
+     model: 'gpt-realtime',
    });
    ```
 

examples/docs/voice-agents/agent.ts

@@ -5,5 +5,5 @@ export const agent = new RealtimeAgent({
 });
 
 export const session = new RealtimeSession(agent, {
-  model: 'gpt-4o-realtime-preview-2025-06-03',
+  model: 'gpt-realtime',
 });

examples/docs/voice-agents/configureSession.ts

@@ -6,7 +6,7 @@ const agent = new RealtimeAgent({
 });
 
 const session = new RealtimeSession(agent, {
-  model: 'gpt-4o-realtime-preview-2025-06-03',
+  model: 'gpt-realtime',
   config: {
     inputAudioFormat: 'pcm16',
     outputAudioFormat: 'pcm16',

examples/docs/voice-agents/createSession.ts

@@ -8,7 +8,7 @@ const agent = new RealtimeAgent({
 async function main() {
   // define which agent you want to start your session with
   const session = new RealtimeSession(agent, {
-    model: 'gpt-4o-realtime-preview-2025-06-03',
+    model: 'gpt-realtime',
   });
   // start your session
   await session.connect({ apiKey: '<your api key>' });

examples/docs/voice-agents/sendMessage.ts

@@ -5,7 +5,7 @@ const agent = new RealtimeAgent({
 });
 
 const session = new RealtimeSession(agent, {
-  model: 'gpt-4o-realtime-preview-2025-06-03',
+  model: 'gpt-realtime',
 });
 
 session.sendMessage('Hello, how are you?');