From 039a09a8e14bac3762cf1c7576e46f5c6a5e5c27 Mon Sep 17 00:00:00 2001
From: skf <54565339+skf-1999@users.noreply.github.com>
Date: Fri, 8 May 2026 21:33:53 +0800
Subject: [PATCH] [Feature] online HunyuanImage-3.0 IT2I (image editing)
 support (#3410)

Signed-off-by: skf1999 <13234016272@163.com>
---
 vllm_omni/entrypoints/openai/api_server.py   | 5 +++++
 vllm_omni/entrypoints/openai/serving_chat.py | 9 +++++++++
 2 files changed, 14 insertions(+)

diff --git a/vllm_omni/entrypoints/openai/api_server.py b/vllm_omni/entrypoints/openai/api_server.py
index 09e4b6b66..e58b490a7 100644
--- a/vllm_omni/entrypoints/openai/api_server.py
+++ b/vllm_omni/entrypoints/openai/api_server.py
@@ -1696,10 +1696,12 @@ async def edit_images(
     # vllm-omni extension for layered models (e.g., Qwen-Image-Layered)
     layers: int | None = Form(None),
     resolution: int | None = Form(None),  # See SUPPORTED_LAYERED_RESOLUTIONS
+    bot_task: str | None = Form(None),
 ) -> ImageGenerationResponse:
     """
     OpenAI-compatible image edit endpoint.
     """
+
     # 1. get engine and model
     engine_client, model_name, stage_configs = _get_engine_and_model(raw_request)
     if model is not None and model != model_name:
@@ -1899,6 +1901,8 @@ async def edit_images(
                 lora_dict = _get_lora_from_json_str(lora)
                 _parse_lora_request(lora_dict)
                 extra_body["lora"] = lora_dict
+            if bot_task is not None:
+                extra_body["bot_task"] = bot_task
 
             prompt_text = prompt.get("prompt", "")
             generation_result = await chat_handler.generate_diffusion_images(
@@ -2218,6 +2222,7 @@ async def _load_input_images(
                 images.append(img)
             except Exception as e:
                 raise ValueError(f"Failed to open uploaded file: {e}")
+
         else:
             raise ValueError(f"Unsupported input: {inp}")
 
diff --git a/vllm_omni/entrypoints/openai/serving_chat.py b/vllm_omni/entrypoints/openai/serving_chat.py
index 7bef559b6..433d16ac7 100644
--- a/vllm_omni/entrypoints/openai/serving_chat.py
+++ b/vllm_omni/entrypoints/openai/serving_chat.py
@@ -2148,6 +2148,7 @@ class OmniOpenAIServingChat(OpenAIServingChat, AudioMixin):
         lora_body = extra_body.get("lora")
         layers = extra_body.get("layers")
         resolution = extra_body.get("resolution")
+        bot_task = extra_body.get("bot_task")
 
         engine_prompt_data: dict[str, Any] | None = None
         modalities = ["image"]
@@ -2158,6 +2159,14 @@ class OmniOpenAIServingChat(OpenAIServingChat, AudioMixin):
             else:
                 engine_prompt_data = {"image": reference_images}
 
+        if bot_task:
+            from vllm_omni.diffusion.models.hunyuan_image3.prompt_utils import build_prompt
+
+            prompt = build_prompt(prompt, task=bot_task)
+            if reference_images and len(reference_images) == 1:
+                engine_prompt_data = {"image": reference_images[0]}
+                modalities = ["image"]
+
         engine_prompt: OmniTextPrompt = {"prompt": prompt}
         engine_prompt["modalities"] = modalities
         if negative_prompt is not None: