Merge pull request #414 from stabey/fix/claude-tool-results-openai-chat

fix: preserve Claude tool results in OpenAI chat conversion

crates/aether-ai-formats/src/formats/conversion/request.rs

@@ -157,7 +157,7 @@ fn request_context(mapped_model: &str, upstream_is_stream: bool) -> FormatContex
 
 #[cfg(test)]
 mod tests {
-    use serde_json::json;
+    use serde_json::{json, Value};
 
     use super::{
         convert_openai_chat_request_to_claude_request,
@@ -215,4 +215,339 @@ mod tests {
         assert_eq!(converted["messages"][0]["role"], "user");
         assert_eq!(converted["messages"][0]["content"], "hello");
     }
+
+    #[test]
+    fn request_normalizer_preserves_multiple_claude_tool_results() {
+        let body = json!({
+            "model": "claude-sonnet",
+            "messages": [
+                {
+                    "role": "assistant",
+                    "content": [
+                        {
+                            "type": "tool_use",
+                            "id": "toolu_1",
+                            "name": "lookup",
+                            "input": {"query": "alpha"}
+                        },
+                        {
+                            "type": "tool_use",
+                            "id": "toolu_2",
+                            "name": "lookup",
+                            "input": {"query": "beta"}
+                        }
+                    ]
+                },
+                {
+                    "role": "user",
+                    "content": [
+                        {
+                            "type": "tool_result",
+                            "tool_use_id": "toolu_1",
+                            "content": "alpha result"
+                        },
+                        {
+                            "type": "tool_result",
+                            "tool_use_id": "toolu_2",
+                            "content": [{"type": "text", "text": "beta result"}]
+                        }
+                    ]
+                }
+            ],
+            "max_tokens": 128,
+        });
+
+        let converted =
+            normalize_claude_request_to_openai_chat_request(&body).expect("openai chat request");
+        let messages = converted["messages"].as_array().expect("messages");
+
+        assert_eq!(messages.len(), 3);
+        assert_eq!(messages[0]["role"], "assistant");
+        assert_eq!(messages[0]["tool_calls"].as_array().unwrap().len(), 2);
+        assert_eq!(messages[0]["tool_calls"][0]["id"], "toolu_1");
+        assert_eq!(messages[0]["tool_calls"][1]["id"], "toolu_2");
+        assert_eq!(messages[1]["role"], "tool");
+        assert_eq!(messages[1]["tool_call_id"], "toolu_1");
+        assert_eq!(messages[1]["content"], "alpha result");
+        assert_eq!(messages[2]["role"], "tool");
+        assert_eq!(messages[2]["tool_call_id"], "toolu_2");
+        assert_eq!(messages[2]["content"], "beta result");
+    }
+
+    #[test]
+    fn request_normalizer_preserves_claude_tool_result_order_around_text() {
+        let body = json!({
+            "model": "claude-sonnet",
+            "messages": [{
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": "before"},
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_1",
+                        "content": "first"
+                    },
+                    {"type": "text", "text": "between"},
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_2",
+                        "content": "second"
+                    }
+                ]
+            }],
+            "max_tokens": 128,
+        });
+
+        let converted =
+            normalize_claude_request_to_openai_chat_request(&body).expect("openai chat request");
+        let messages = converted["messages"].as_array().expect("messages");
+
+        assert_eq!(messages.len(), 4);
+        assert_eq!(messages[0]["role"], "user");
+        assert_eq!(messages[0]["content"], "before");
+        assert_eq!(messages[1]["role"], "tool");
+        assert_eq!(messages[1]["tool_call_id"], "toolu_1");
+        assert_eq!(messages[1]["content"], "first");
+        assert_eq!(messages[2]["role"], "user");
+        assert_eq!(messages[2]["content"], "between");
+        assert_eq!(messages[3]["role"], "tool");
+        assert_eq!(messages[3]["tool_call_id"], "toolu_2");
+        assert_eq!(messages[3]["content"], "second");
+    }
+
+    #[test]
+    fn request_normalizer_marks_claude_error_tool_result_string_and_object_content() {
+        let object_result = json!({"code": "ENOENT", "message": "missing"});
+        let body = json!({
+            "model": "claude-sonnet",
+            "messages": [{
+                "role": "user",
+                "content": [
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_error_string",
+                        "content": "lookup failed",
+                        "is_error": true
+                    },
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_error_empty",
+                        "content": "",
+                        "is_error": true
+                    },
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_error_object",
+                        "content": object_result,
+                        "is_error": true
+                    },
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_ok",
+                        "content": "still ok"
+                    }
+                ]
+            }],
+            "max_tokens": 128,
+        });
+
+        let converted =
+            normalize_claude_request_to_openai_chat_request(&body).expect("openai chat request");
+        let messages = converted["messages"].as_array().expect("messages");
+
+        assert_eq!(messages.len(), 4);
+        assert_eq!(messages[0]["role"], "tool");
+        assert_eq!(messages[0]["tool_call_id"], "toolu_error_string");
+        assert_eq!(messages[0]["content"], "[tool error]\nlookup failed");
+
+        assert_eq!(messages[1]["role"], "tool");
+        assert_eq!(messages[1]["tool_call_id"], "toolu_error_empty");
+        assert_eq!(messages[1]["content"], "[tool error]");
+
+        assert_eq!(messages[2]["role"], "tool");
+        assert_eq!(messages[2]["tool_call_id"], "toolu_error_object");
+        let object_content = messages[2]["content"].as_str().expect("object content");
+        let serialized_object = object_content
+            .strip_prefix("[tool error]\n")
+            .expect("error prefix");
+        assert_eq!(
+            serde_json::from_str::<Value>(serialized_object).expect("serialized object"),
+            object_result
+        );
+
+        assert_eq!(messages[3]["role"], "tool");
+        assert_eq!(messages[3]["tool_call_id"], "toolu_ok");
+        assert_eq!(messages[3]["content"], "still ok");
+    }
+
+    #[test]
+    fn request_normalizer_marks_claude_error_tool_result_multipart_image_content() {
+        let body = json!({
+            "model": "claude-sonnet",
+            "messages": [{
+                "role": "user",
+                "content": [{
+                    "type": "tool_result",
+                    "tool_use_id": "toolu_error_image",
+                    "content": [
+                        {"type": "text", "text": "preview"},
+                        {
+                            "type": "image",
+                            "source": {
+                                "type": "base64",
+                                "media_type": "image/png",
+                                "data": "aW1hZ2U="
+                            }
+                        }
+                    ],
+                    "is_error": true
+                }]
+            }],
+            "max_tokens": 128,
+        });
+
+        let converted =
+            normalize_claude_request_to_openai_chat_request(&body).expect("openai chat request");
+        let messages = converted["messages"].as_array().expect("messages");
+
+        assert_eq!(messages.len(), 1);
+        assert_eq!(messages[0]["role"], "tool");
+        assert_eq!(messages[0]["tool_call_id"], "toolu_error_image");
+        let content = messages[0]["content"]
+            .as_array()
+            .expect("multipart error content");
+        assert_eq!(
+            content.as_slice(),
+            &[
+                json!({"type": "text", "text": "[tool error]"}),
+                json!({"type": "text", "text": "preview"}),
+                json!({
+                    "type": "image_url",
+                    "image_url": {"url": "data:image/png;base64,aW1hZ2U="}
+                }),
+            ]
+        );
+    }
+
+    #[test]
+    fn request_normalizer_preserves_legal_openai_tool_content_for_claude_variants() {
+        let anthropic_blocks = json!([
+            {"type": "text", "text": "preview"},
+            {
+                "type": "image",
+                "source": {
+                    "type": "base64",
+                    "media_type": "image/jpeg",
+                    "data": "aGVsbG8="
+                }
+            },
+            {
+                "type": "image",
+                "source": {
+                    "type": "url",
+                    "url": "https://example.com/image.jpg"
+                }
+            },
+            {
+                "type": "document",
+                "source": {
+                    "type": "base64",
+                    "media_type": "application/pdf",
+                    "data": "JVBERi0x"
+                }
+            },
+            {
+                "type": "document",
+                "source": {
+                    "type": "url",
+                    "url": "https://example.com/report.pdf"
+                }
+            },
+            {
+                "type": "document",
+                "source": {
+                    "type": "text",
+                    "media_type": "text/plain",
+                    "data": "document body"
+                }
+            }
+        ]);
+        let object_result = json!({"answer": 42, "ok": true});
+        let body = json!({
+            "model": "claude-sonnet",
+            "messages": [{
+                "role": "user",
+                "content": [
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_object",
+                        "content": object_result
+                    },
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_text_blocks",
+                        "content": [
+                            {"type": "text", "text": "line one"},
+                            {"type": "text", "text": "line two"}
+                        ]
+                    },
+                    {
+                        "type": "tool_result",
+                        "tool_use_id": "toolu_anthropic_blocks",
+                        "content": anthropic_blocks
+                    }
+                ]
+            }],
+            "max_tokens": 128,
+        });
+
+        let converted =
+            normalize_claude_request_to_openai_chat_request(&body).expect("openai chat request");
+        let messages = converted["messages"].as_array().expect("messages");
+
+        assert_eq!(messages.len(), 3);
+        assert_eq!(messages[0]["role"], "tool");
+        assert_eq!(messages[0]["tool_call_id"], "toolu_object");
+        let object_content = messages[0]["content"].as_str().expect("object content");
+        assert_eq!(
+            serde_json::from_str::<Value>(object_content).expect("serialized object"),
+            object_result
+        );
+
+        assert_eq!(messages[1]["role"], "tool");
+        assert_eq!(messages[1]["tool_call_id"], "toolu_text_blocks");
+        assert_eq!(messages[1]["content"], "line one\n\nline two");
+
+        assert_eq!(messages[2]["role"], "tool");
+        assert_eq!(messages[2]["tool_call_id"], "toolu_anthropic_blocks");
+        let block_content = messages[2]["content"]
+            .as_array()
+            .expect("multipart anthropic block content");
+        assert_eq!(
+            block_content.as_slice(),
+            &[
+                json!({"type": "text", "text": "preview"}),
+                json!({
+                    "type": "image_url",
+                    "image_url": {"url": "data:image/jpeg;base64,aGVsbG8="}
+                }),
+                json!({
+                    "type": "image_url",
+                    "image_url": {"url": "https://example.com/image.jpg"}
+                }),
+                json!({
+                    "type": "file",
+                    "file": {"file_data": "data:application/pdf;base64,JVBERi0x"}
+                }),
+                json!({"type": "text", "text": "[File: https://example.com/report.pdf]"}),
+                json!({
+                    "type": "text",
+                    "text": "[Claude tool_result document content omitted: text/plain]"
+                }),
+            ]
+        );
+        let block_content_json = Value::Array(block_content.clone()).to_string();
+        assert!(!block_content_json.contains("\"source\""));
+        assert!(!block_content_json.contains("document body"));
+    }
 }

crates/aether-ai-formats/src/formats/openai/chat/request.rs

@@ -3,7 +3,7 @@ use serde_json::{json, Value};
 use crate::{
     formats::context::FormatContext,
     protocol::canonical::{
-        canonical_extension_object_mut, canonical_message_to_openai_chat,
+        canonical_extension_object_mut, canonical_message_to_openai_chat_messages,
         canonical_response_format_to_openai, canonical_tool_choice_to_openai,
         canonical_tool_to_openai, namespace_extension_object, openai_content_text,
         openai_extensions, openai_generation_config, openai_message_content_blocks,
@@ -148,7 +148,7 @@ pub fn to_raw(canonical: &CanonicalRequest) -> Value {
         }
     }
     for message in &canonical.messages {
-        messages.push(canonical_message_to_openai_chat(message));
+        messages.extend(canonical_message_to_openai_chat_messages(message));
     }
     output.insert("messages".to_string(), Value::Array(messages));
 

crates/aether-ai-formats/src/protocol/canonical.rs

@@ -9,6 +9,9 @@ pub use crate::protocol::stream::{CanonicalStreamEvent, CanonicalStreamFrame};
 
 pub(crate) const OPENAI_RESPONSES_EXTENSION_NAMESPACE: &str = "openai_responses";
 pub(crate) const OPENAI_RESPONSES_LEGACY_EXTENSION_NAMESPACE: &str = "openai_cli";
+const AETHER_EXTENSION_NAMESPACE: &str = "aether";
+const CLAUDE_TOOL_RESULT_SOURCE_MARKER: &str = "claude_tool_result";
+const OPENAI_CHAT_TOOL_ERROR_PREFIX: &str = "[tool error]";
 
 #[derive(Debug, Clone, Default, PartialEq, Eq, Serialize, Deserialize)]
 #[serde(rename_all = "snake_case")]
@@ -1191,6 +1194,14 @@ pub(crate) fn claude_block_to_canonical_block(block: &Value) -> Option<Canonical
         }),
         "tool_result" => {
             let content = block_object.get("content").cloned();
+            let mut extensions = claude_extensions(
+                block_object,
+                &["type", "tool_use_id", "content", "is_error"],
+            );
+            extensions.insert(
+                AETHER_EXTENSION_NAMESPACE.to_string(),
+                json!({ "source": CLAUDE_TOOL_RESULT_SOURCE_MARKER }),
+            );
             Some(CanonicalContentBlock::ToolResult {
                 tool_use_id: block_object
                     .get("tool_use_id")
@@ -1204,10 +1215,7 @@ pub(crate) fn claude_block_to_canonical_block(block: &Value) -> Option<Canonical
                     .get("is_error")
                     .and_then(Value::as_bool)
                     .unwrap_or(false),
-                extensions: claude_extensions(
-                    block_object,
-                    &["type", "tool_use_id", "content", "is_error"],
-                ),
+                extensions,
             })
         }
         _ => Some(CanonicalContentBlock::Unknown {
@@ -2050,7 +2058,64 @@ pub(crate) fn openai_part_to_canonical_block(part: &Value) -> Option<CanonicalCo
     }
 }
 
-pub(crate) fn canonical_message_to_openai_chat(message: &CanonicalMessage) -> Value {
+pub(crate) fn canonical_message_to_openai_chat_messages(message: &CanonicalMessage) -> Vec<Value> {
+    let mut messages = Vec::new();
+    let mut pending_start = 0usize;
+    let mut saw_tool_result = false;
+
+    for (index, block) in message.content.iter().enumerate() {
+        if let CanonicalContentBlock::ToolResult { .. } = block {
+            saw_tool_result = true;
+            if pending_start < index {
+                if let Some(message_value) = canonical_message_blocks_to_openai_chat(
+                    message,
+                    &message.content[pending_start..index],
+                    false,
+                ) {
+                    messages.push(message_value);
+                }
+            }
+            messages.push(canonical_tool_result_to_openai_chat(block));
+            pending_start = index + 1;
+        }
+    }
+
+    if !saw_tool_result {
+        return vec![canonical_message_without_tool_results_to_openai_chat(
+            message,
+        )];
+    }
+
+    if pending_start < message.content.len() {
+        if let Some(message_value) = canonical_message_blocks_to_openai_chat(
+            message,
+            &message.content[pending_start..],
+            false,
+        ) {
+            messages.push(message_value);
+        }
+    }
+
+    messages
+}
+
+fn canonical_message_without_tool_results_to_openai_chat(message: &CanonicalMessage) -> Value {
+    debug_assert!(
+        !message
+            .content
+            .iter()
+            .any(|block| matches!(block, CanonicalContentBlock::ToolResult { .. })),
+        "single OpenAI Chat message emission requires no ToolResult blocks; use canonical_message_to_openai_chat_messages"
+    );
+    canonical_message_blocks_to_openai_chat(message, &message.content, true)
+        .expect("include_empty=true always emits a chat message")
+}
+
+fn canonical_message_blocks_to_openai_chat(
+    message: &CanonicalMessage,
+    content: &[CanonicalContentBlock],
+    include_empty: bool,
+) -> Option<Value> {
     let mut output = Map::new();
     output.insert(
         "role".to_string(),
@@ -2069,7 +2134,7 @@ pub(crate) fn canonical_message_to_openai_chat(message: &CanonicalMessage) -> Va
     let mut tool_calls = Vec::new();
     let mut reasoning_segments = Vec::new();
     let mut reasoning_parts = Vec::new();
-    for block in &message.content {
+    for block in content {
         match block {
             CanonicalContentBlock::Thinking {
                 text,
@@ -2123,25 +2188,7 @@ pub(crate) fn canonical_message_to_openai_chat(message: &CanonicalMessage) -> Va
                     "arguments": canonicalize_tool_arguments(input),
                 }
             })),
-            CanonicalContentBlock::ToolResult {
-                tool_use_id,
-                content_text,
-                output: result_output,
-                ..
-            } => {
-                output.insert("role".to_string(), Value::String("tool".to_string()));
-                output.insert(
-                    "tool_call_id".to_string(),
-                    Value::String(tool_use_id.clone()),
-                );
-                output.insert(
-                    "content".to_string(),
-                    result_output
-                        .clone()
-                        .unwrap_or_else(|| Value::String(content_text.clone().unwrap_or_default())),
-                );
-                return Value::Object(output);
-            }
+            CanonicalContentBlock::ToolResult { .. } => {}
             other => {
                 if let Some(part) = canonical_content_block_to_openai_part(other) {
                     content_parts.push(part);
@@ -2149,6 +2196,14 @@ pub(crate) fn canonical_message_to_openai_chat(message: &CanonicalMessage) -> Va
             }
         }
     }
+    if !include_empty
+        && content_parts.is_empty()
+        && tool_calls.is_empty()
+        && reasoning_segments.is_empty()
+        && reasoning_parts.is_empty()
+    {
+        return None;
+    }
     output.insert(
         "content".to_string(),
         if !tool_calls.is_empty() && content_parts.is_empty() {
@@ -2173,9 +2228,254 @@ pub(crate) fn canonical_message_to_openai_chat(message: &CanonicalMessage) -> Va
     if !reasoning_parts.is_empty() {
         output.insert("reasoning_parts".to_string(), Value::Array(reasoning_parts));
     }
+    Some(Value::Object(output))
+}
+
+fn canonical_tool_result_to_openai_chat(block: &CanonicalContentBlock) -> Value {
+    let CanonicalContentBlock::ToolResult {
+        tool_use_id,
+        content_text,
+        output: result_output,
+        is_error,
+        extensions,
+        ..
+    } = block
+    else {
+        unreachable!("canonical_tool_result_to_openai_chat requires ToolResult");
+    };
+
+    let mut output = Map::new();
+    output.insert("role".to_string(), Value::String("tool".to_string()));
+    output.insert(
+        "tool_call_id".to_string(),
+        Value::String(tool_use_id.clone()),
+    );
+    let content = if is_claude_tool_result(extensions) {
+        let content =
+            openai_chat_tool_result_content(result_output.as_ref(), content_text.as_deref());
+        if *is_error {
+            openai_chat_tool_error_content(content)
+        } else {
+            content
+        }
+    } else {
+        result_output
+            .clone()
+            .unwrap_or_else(|| Value::String(content_text.clone().unwrap_or_default()))
+    };
+    output.insert("content".to_string(), content);
     Value::Object(output)
 }
 
+fn is_claude_tool_result(extensions: &BTreeMap<String, Value>) -> bool {
+    extensions
+        .get(AETHER_EXTENSION_NAMESPACE)
+        .and_then(|value| value.get("source"))
+        .and_then(Value::as_str)
+        == Some(CLAUDE_TOOL_RESULT_SOURCE_MARKER)
+}
+
+fn openai_chat_tool_result_content(output: Option<&Value>, content_text: Option<&str>) -> Value {
+    match output {
+        Some(Value::String(text)) => Value::String(text.clone()),
+        Some(Value::Array(parts)) => anthropic_tool_result_blocks_to_openai_chat_content(parts),
+        Some(value) => Value::String(value.to_string()),
+        None => Value::String(content_text.unwrap_or_default().to_string()),
+    }
+}
+
+fn openai_chat_tool_error_content(content: Value) -> Value {
+    match content {
+        Value::String(text) if text.is_empty() => {
+            Value::String(OPENAI_CHAT_TOOL_ERROR_PREFIX.to_string())
+        }
+        Value::String(text) => Value::String(format!("{OPENAI_CHAT_TOOL_ERROR_PREFIX}\n{text}")),
+        Value::Array(parts) => {
+            let mut prefixed_parts = Vec::with_capacity(parts.len() + 1);
+            prefixed_parts.push(openai_text_part(OPENAI_CHAT_TOOL_ERROR_PREFIX));
+            prefixed_parts.extend(parts);
+            Value::Array(prefixed_parts)
+        }
+        value => Value::String(format!("{OPENAI_CHAT_TOOL_ERROR_PREFIX}\n{value}")),
+    }
+}
+
+fn anthropic_tool_result_blocks_to_openai_chat_content(parts: &[Value]) -> Value {
+    if let Some(text) = anthropic_text_blocks_to_string(parts) {
+        return Value::String(text);
+    }
+
+    let mut has_media_part = false;
+    let converted_parts = parts
+        .iter()
+        .map(|part| {
+            let openai_part = anthropic_tool_result_block_to_openai_chat_part(part);
+            if !openai_chat_part_is_text(&openai_part) {
+                has_media_part = true;
+            }
+            openai_part
+        })
+        .collect::<Vec<_>>();
+
+    if has_media_part {
+        Value::Array(converted_parts)
+    } else {
+        Value::String(openai_text_parts_to_string(&converted_parts))
+    }
+}
+
+fn anthropic_text_blocks_to_string(parts: &[Value]) -> Option<String> {
+    let mut texts = Vec::with_capacity(parts.len());
+    for part in parts {
+        let part_object = part.as_object()?;
+        if part_object.get("type").and_then(Value::as_str) != Some("text") {
+            return None;
+        }
+        texts.push(part_object.get("text").and_then(Value::as_str)?);
+    }
+    Some(texts.join("\n\n"))
+}
+
+fn anthropic_tool_result_block_to_openai_chat_part(part: &Value) -> Value {
+    let Some(part_object) = part.as_object() else {
+        return openai_text_part("[Claude tool_result non-text content omitted]");
+    };
+    match part_object
+        .get("type")
+        .and_then(Value::as_str)
+        .unwrap_or_default()
+    {
+        "text" => openai_text_part(
+            part_object
+                .get("text")
+                .and_then(Value::as_str)
+                .unwrap_or_default(),
+        ),
+        "image" => anthropic_image_block_to_openai_chat_part(part_object).unwrap_or_else(|| {
+            openai_text_part(anthropic_media_block_summary("image", part_object))
+        }),
+        "document" => {
+            anthropic_document_block_to_openai_chat_part(part_object).unwrap_or_else(|| {
+                openai_text_part(anthropic_media_block_summary("document", part_object))
+            })
+        }
+        "file" => anthropic_document_block_to_openai_chat_part(part_object).unwrap_or_else(|| {
+            openai_text_part(anthropic_media_block_summary("file", part_object))
+        }),
+        "" => openai_text_part("[Claude tool_result object content omitted]"),
+        raw_type => openai_text_part(format!("[Claude tool_result {raw_type} content omitted]")),
+    }
+}
+
+fn anthropic_image_block_to_openai_chat_part(block: &Map<String, Value>) -> Option<Value> {
+    let source = block.get("source")?.as_object()?;
+    match source
+        .get("type")
+        .and_then(Value::as_str)
+        .unwrap_or_default()
+    {
+        "base64" => {
+            let media_type = anthropic_source_media_type(source)?;
+            let data = anthropic_source_str(source, "data")?;
+            Some(json!({
+                "type": "image_url",
+                "image_url": {
+                    "url": format!("data:{media_type};base64,{data}"),
+                },
+            }))
+        }
+        "url" => {
+            let url = anthropic_source_str(source, "url")?;
+            Some(json!({
+                "type": "image_url",
+                "image_url": {
+                    "url": url,
+                },
+            }))
+        }
+        _ => None,
+    }
+}
+
+fn anthropic_document_block_to_openai_chat_part(block: &Map<String, Value>) -> Option<Value> {
+    let source = block.get("source")?.as_object()?;
+    match source
+        .get("type")
+        .and_then(Value::as_str)
+        .unwrap_or_default()
+    {
+        "base64" => {
+            let media_type = anthropic_source_media_type(source)?;
+            let data = anthropic_source_str(source, "data")?;
+            Some(json!({
+                "type": "file",
+                "file": {
+                    "file_data": format!("data:{media_type};base64,{data}"),
+                },
+            }))
+        }
+        "url" => {
+            let url = anthropic_source_str(source, "url")?;
+            Some(openai_text_part(format!("[File: {url}]")))
+        }
+        _ => None,
+    }
+}
+
+fn anthropic_media_block_summary(kind: &str, block: &Map<String, Value>) -> String {
+    let media_type = block
+        .get("source")
+        .and_then(Value::as_object)
+        .and_then(anthropic_source_media_type);
+    match media_type {
+        Some(media_type) if !media_type.trim().is_empty() => {
+            format!("[Claude tool_result {kind} content omitted: {media_type}]")
+        }
+        _ => format!("[Claude tool_result {kind} content omitted]"),
+    }
+}
+
+fn openai_text_part(text: impl Into<String>) -> Value {
+    json!({
+        "type": "text",
+        "text": text.into(),
+    })
+}
+
+fn openai_chat_part_is_text(part: &Value) -> bool {
+    part.as_object()
+        .and_then(|object| object.get("type"))
+        .and_then(Value::as_str)
+        == Some("text")
+}
+
+fn openai_text_parts_to_string(parts: &[Value]) -> String {
+    parts
+        .iter()
+        .filter_map(|part| {
+            part.as_object()
+                .and_then(|object| object.get("text"))
+                .and_then(Value::as_str)
+        })
+        .collect::<Vec<_>>()
+        .join("\n\n")
+}
+
+fn anthropic_source_media_type(source: &Map<String, Value>) -> Option<&str> {
+    source
+        .get("media_type")
+        .or_else(|| source.get("mime_type"))
+        .and_then(Value::as_str)
+        .filter(|value| !value.trim().is_empty())
+}
+
+fn anthropic_source_str<'a>(source: &'a Map<String, Value>, key: &str) -> Option<&'a str> {
+    source
+        .get(key)
+        .and_then(Value::as_str)
+        .filter(|value| !value.trim().is_empty())
+}
+
 pub(crate) fn canonical_content_block_to_openai_part(
     block: &CanonicalContentBlock,
 ) -> Option<Value> {