Gate audio history by model input modalities (#33982)

## What changed

- Add `audio` to the model input modalities exposed by the protocol and generated app-server schemas.
- Preserve audio in prompts for models that advertise audio input, and replace historical audio with an omission marker for models that do not.

## Testing

- Cover prompt normalization with and without audio support.
- Extend the model-switching integration test to verify that switching from a multimodal model to a text-only model strips prior image and audio content.

GitOrigin-RevId: c18e47ad602720216303ff9abb5bca4d508391ee

codex-rs/core/src/context_manager/history.rs

@@ -135,9 +135,8 @@ impl ContextManager {
     }
 
     /// Returns the history prepared for sending to the model. This applies a proper
-    /// normalization and drops un-suited items. When `input_modalities` does not
-    /// include `InputModality::Image`, images are stripped from messages and tool
-    /// outputs.
+    /// normalization and drops un-suited items. Unsupported image and audio content
+    /// is stripped from messages and tool outputs according to `input_modalities`.
     pub(crate) fn for_prompt(mut self, input_modalities: &[InputModality]) -> Vec<ResponseItem> {
         self.normalize_history(input_modalities);
         self.items
@@ -355,7 +354,7 @@ impl ContextManager {
     /// This function enforces a couple of invariants on the in-memory history:
     /// 1. every call (function/custom) has a corresponding output entry
     /// 2. every output has a corresponding call entry
-    /// 3. when images are unsupported, image content is stripped from messages and tool outputs
+    /// 3. unsupported image and audio content is stripped from messages and tool outputs
     fn normalize_history(&mut self, input_modalities: &[InputModality]) {
         // all function/tool calls must have a corresponding output
         normalize::ensure_call_outputs_present(&mut self.items);
@@ -365,6 +364,9 @@ impl ContextManager {
 
         // strip images when model does not support them
         normalize::strip_images_when_unsupported(input_modalities, &mut self.items);
+
+        // strip audio when model does not support it
+        normalize::strip_audio_when_unsupported(input_modalities, &mut self.items);
     }
 
     fn process_item(&self, item: &ResponseItem, policy: TruncationPolicy) -> ResponseItem {

codex-rs/core/src/context_manager/history_tests.rs

@@ -471,7 +471,7 @@ fn total_token_usage_includes_all_items_after_last_model_generated_item() {
 }
 
 #[test]
-fn for_prompt_strips_images_when_model_does_not_support_images() {
+fn for_prompt_strips_media_when_model_does_not_support_it() {
     let items = vec![
         ResponseItem::Message {
             id: None,
@@ -484,6 +484,9 @@ fn for_prompt_strips_images_when_model_does_not_support_images() {
                     image_url: "https://example.com/img.png".to_string(),
                     detail: Some(DEFAULT_IMAGE_DETAIL),
                 },
+                ContentItem::InputAudio {
+                    audio_url: "data:audio/wav;base64,YXVkaW8=".to_string(),
+                },
                 ContentItem::InputText {
                     text: "caption".to_string(),
                 },
@@ -554,6 +557,10 @@ fn for_prompt_strips_images_when_model_does_not_support_images() {
                     text: "image content omitted because you do not support image input"
                         .to_string(),
                 },
+                ContentItem::InputText {
+                    text: "audio content omitted because you do not support audio input"
+                        .to_string(),
+                },
                 ContentItem::InputText {
                     text: "caption".to_string(),
                 },
@@ -635,6 +642,21 @@ fn for_prompt_strips_images_when_model_does_not_support_images() {
     } else {
         panic!("expected Message");
     }
+
+    let audio_message = ResponseItem::Message {
+        id: None,
+        role: "user".to_string(),
+        content: vec![ContentItem::InputAudio {
+            audio_url: "data:audio/wav;base64,YXVkaW8=".to_string(),
+        }],
+        phase: None,
+        internal_chat_message_metadata_passthrough: None,
+    };
+    let with_audio = create_history_with_items(vec![audio_message.clone()]);
+    assert_eq!(
+        with_audio.for_prompt(&[InputModality::Text, InputModality::Audio]),
+        vec![audio_message]
+    );
 }
 
 #[test]

codex-rs/core/src/context_manager/normalize.rs

@@ -12,6 +12,8 @@ use tracing::info;
 
 const IMAGE_CONTENT_OMITTED_PLACEHOLDER: &str =
     "image content omitted because you do not support image input";
+const AUDIO_CONTENT_OMITTED_PLACEHOLDER: &str =
+    "audio content omitted because you do not support audio input";
 // Changing this value would change model-visible IDs and invalidate prompt caches.
 const SYNTHETIC_OUTPUT_ID_NAMESPACE: Uuid = Uuid::from_u128(0x90d38d3e_6a5b_4d52_bfe2_2f1e634bfac4);
 
@@ -366,3 +368,26 @@ pub(crate) fn strip_images_when_unsupported(
         }
     }
 }
+
+/// Strip audio content from messages when the model does not support audio.
+/// When `input_modalities` contains `InputModality::Audio`, no stripping is performed.
+pub(crate) fn strip_audio_when_unsupported(
+    input_modalities: &[InputModality],
+    items: &mut [ResponseItem],
+) {
+    if input_modalities.contains(&InputModality::Audio) {
+        return;
+    }
+
+    for item in items.iter_mut() {
+        if let ResponseItem::Message { content, .. } = item {
+            for content_item in content.iter_mut() {
+                if matches!(content_item, ContentItem::InputAudio { .. }) {
+                    *content_item = ContentItem::InputText {
+                        text: AUDIO_CONTENT_OMITTED_PLACEHOLDER.to_string(),
+                    };
+                }
+            }
+        }
+    }
+}