fix: Fix Python code and add multi-language support to Anthropic message-per-token guide

- Add multi-language support (Python, Java, Swift) to message-per-token
  guide, matching the structure of message-per-response
- Fix Python code in both guides: use AsyncAnthropic with async/await,
  use Message objects for publish with extras and append_message,
  use transport_params for echo suppression
- Scope "without await" asides to JavaScript only

Author: GregHolmes

src/pages/docs/guides/ai-transport/anthropic-message-per-response.mdx

@@ -229,28 +229,29 @@ streamAnthropicResponse("Tell me a short joke");
 ```
 
 ```agent_python
+import asyncio
 import anthropic
 
 # Initialize Anthropic client
-client = anthropic.Anthropic()
+client = anthropic.AsyncAnthropic()
 
 # Process each streaming event
-def process_event(event):
+async def process_event(event):
     print(event)
     # This function is updated in the next sections
 
 # Create streaming response from Anthropic
-def stream_anthropic_response(prompt: str):
-    with client.messages.stream(
+async def stream_anthropic_response(prompt: str):
+    async with client.messages.stream(
         model="claude-sonnet-4-5",
         max_tokens=1024,
         messages=[{"role": "user", "content": prompt}],
     ) as stream:
-        for event in stream:
-            process_event(event)
+        async for event in stream:
+            await process_event(event)
 
 # Usage example
-stream_anthropic_response("Tell me a short joke")
+asyncio.run(stream_anthropic_response("Tell me a short joke"))
 ```
 
 ```agent_java
@@ -364,7 +365,7 @@ const channel = realtime.channels.get('ai:{{RANDOM_CHANNEL_NAME}}');
 from ably import AblyRealtime
 
 # Initialize Ably Realtime client
-realtime = AblyRealtime(key='{{API_KEY}}', echo_messages=False)
+realtime = AblyRealtime(key='{{API_KEY}}', transport_params={'echo': 'false'})
 
 # Create a channel for publishing streamed AI responses
 channel = realtime.channels.get('ai:{{RANDOM_CHANNEL_NAME}}')
@@ -446,6 +447,8 @@ async function processEvent(event) {
 ```
 
 ```agent_python
+from ably.types.message import Message
+
 # Track state across events
 msg_serial = None
 text_block_index = None
@@ -456,7 +459,7 @@ async def process_event(event):
 
     if event.type == 'message_start':
         # Publish initial empty message when response starts
-        result = await channel.publish('response', data='')
+        result = await channel.publish('response', '')
 
         # Capture the message serial for appending tokens
         msg_serial = result.serials[0]
@@ -471,7 +474,9 @@ async def process_event(event):
         if (event.index == text_block_index and
             hasattr(event.delta, 'text') and
             msg_serial):
-            channel.append_message(serial=msg_serial, data=event.delta.text)
+            await channel.append_message(
+                Message(serial=msg_serial, data=event.delta.text)
+            )
 
     elif event.type == 'message_stop':
         print('Stream completed!')
@@ -529,9 +534,11 @@ This implementation:
 - Filters for `content_block_delta` events with `text_delta` type from text content blocks
 - Appends each token to the original message
 
+<If agent_lang="javascript">
 <Aside data-type="note">
 Append operations are published without `await` to maximize throughput. Ably maintains message ordering even without awaiting each append. For more information, see [Publishing tokens](/docs/ai-transport/token-streaming/message-per-response#publishing).
 </Aside>
+</If>
 
 <Aside data-type="important">
 Standard Ably message [size limits](/docs/platform/pricing/limits#message) apply to the complete concatenated message. If appending a token would exceed the maximum message size, the append is rejected.
@@ -732,7 +739,9 @@ Subscribers receive different message actions depending on when they join and ho
 
 - `message.update`: Contains the whole response up to that point. The message `data` contains the full concatenated text so far. Replace the entire response content with this data for the message identified by `serial`. This action occurs when the channel needs to resynchronize the full message state, such as after a client [resumes](/docs/connect/states#resume) from a transient disconnection.
 
+<If client_lang="javascript,java">
 Run the subscriber in a separate terminal:
+</If>
 
 <If client_lang="javascript">
 <Code>
@@ -755,7 +764,12 @@ mvn compile exec:java -Dexec.mainClass="Subscriber"
 </Code>
 </If>
 
+<If client_lang="javascript,java">
 With the subscriber running, run the publisher in another terminal. The tokens stream in realtime as the Anthropic model generates them.
+</If>
+<If client_lang="swift">
+With the subscriber running, run the publisher in a terminal. The tokens stream in realtime as the Anthropic model generates them.
+</If>
 
 ## Step 5: Stream with multiple publishers and subscribers <a id="step-5"/>
 
@@ -765,7 +779,9 @@ Ably's [channel-oriented sessions](/docs/ai-transport/sessions-identity#connecti
 
 Each subscriber receives the complete stream of tokens independently, enabling you to build collaborative experiences or multi-device applications.
 
+<If client_lang="javascript,java">
 Run a subscriber in multiple separate terminals:
+</If>
 
 <If client_lang="javascript">
 <Code>