diff --git a/docs/.nav.yml b/docs/.nav.yml
index 79d7c38e274..455a0525056 100644
--- a/docs/.nav.yml
+++ b/docs/.nav.yml
@@ -107,7 +107,7 @@ nav:
       - design/feature/hsdp.md
       - design/feature/cache_dit.md
       - design/feature/teacache.md
-      - design/feature/async_chunk_design.md
+      - design/feature/async_chunk.md
       - design/feature/vae_parallel.md
       - design/feature/diffusion_step_execution.md
     - Module Design:
diff --git a/docs/contributing/model/adding_omni_model.md b/docs/contributing/model/adding_omni_model.md
index 478e77c7d50..1eaff10596c 100644
--- a/docs/contributing/model/adding_omni_model.md
+++ b/docs/contributing/model/adding_omni_model.md
@@ -408,18 +408,17 @@ Understanding the data structures is crucial for implementing stage transitions:
 
 **Input to your function:**
 - `stage_list[source_stage_id].engine_outputs`: List of `EngineCoreOutput` objects
-  - Each contains `outputs`: List of `RequestOutput` objects
-  - Each `RequestOutput` has:
-    - `token_ids`: Generated token IDs
-    - `multimodal_output`: Dict with keys like `"code_predictor_codes"`, etc.
-      - These are the hidden states or intermediate outputs from the model's forward pass
-    - `prompt_token_ids`: Original prompt token IDs
+-   -  Each contains `outputs`: List of `RequestOutput` objects
+    - Each `RequestOutput` has:
+-   -  - `token_ids`: Generated token IDs
+       - `multimodal_output`: Dict with keys like `"code_predictor_codes"`, etc.These are the hidden states or intermediate outputs from the model's forward pass
+       - `prompt_token_ids`: Original prompt token IDs
 
 **Output from your function:**
 - Must return `list[OmniTokensPrompt]` where each `OmniTokensPrompt` contains:
-  - `prompt_token_ids`: List[int] - Token IDs for the next stage
-  - `additional_information`: Dict[str, Any] - Optional metadata (e.g., embeddings, hidden states)
-  - `multi_modal_data`: Optional multimodal data if needed
+-   - `prompt_token_ids`: List[int] - Token IDs for the next stage
+    - `additional_information`: Dict[str, Any] - Optional metadata (e.g., embeddings, hidden states)
+    - `multi_modal_data`: Optional multimodal data if needed
 
 ### How Model Outputs Are Stored
 
diff --git a/docs/contributing/model/adding_tts_model.md b/docs/contributing/model/adding_tts_model.md
index 66da1749cee..622064173cd 100644
--- a/docs/contributing/model/adding_tts_model.md
+++ b/docs/contributing/model/adding_tts_model.md
@@ -28,7 +28,7 @@ and can be placed on different devices. Qwen3-TTS has two stages:
 
 Each stage is a separate model class configured independently via YAML. The two stages
 are connected by the `async_chunk` framework, which enables inter-stage streaming for
-low first-packet latency (see [Async Chunk Design](../../design/feature/async_chunk_design.md)).
+low first-packet latency (see [Async Chunk Design](../../design/feature/async_chunk.md)).
 
 ### Without async_chunk (batch mode)
 
@@ -591,5 +591,5 @@ Adding a TTS model to vLLM-Omni involves:
 For more information, see:
 
 - [Architecture Overview](../../design/architecture_overview.md)
-- [Async Chunk Design](../../design/feature/async_chunk_design.md)
+- [Async Chunk Design](../../design/feature/async_chunk.md)
 - [Stage Configuration Guide](../../configuration/stage_configs.md)
diff --git a/docs/design/feature/async_chunk_design.md b/docs/design/feature/async_chunk.md
similarity index 99%
rename from docs/design/feature/async_chunk_design.md
rename to docs/design/feature/async_chunk.md
index 45314a0aec6..57b4209b8df 100644
--- a/docs/design/feature/async_chunk_design.md
+++ b/docs/design/feature/async_chunk.md
@@ -1,4 +1,4 @@
-# Async Chunk Design
+# Async Chunk
 
 ## Table of Contents
 
@@ -88,8 +88,9 @@ The following diagram illustrates the **Async Chunk Architecture** for multi-sta
 </p>
 
 **Diagram Legend:**
+
 | Step | Stage Type | Description |
-|:------:|:-----------:|:------------|
+|------|-----------|------------|
 | `prefill` | Initialization | Context processing, KV cache initialization |
 | `decode` | Autoregressive | Token-by-token generation in AR stages |
 | `codes` | Audio Encoding | RVQ codec codes from Talker stage |
diff --git a/docs/source/architecture/async-chunk-architecture.png b/docs/source/architecture/async-chunk-architecture.png
index 249de53bfe3..7b3e95e4df9 100644
Binary files a/docs/source/architecture/async-chunk-architecture.png and b/docs/source/architecture/async-chunk-architecture.png differ
diff --git a/docs/source/architecture/vllm-omni-dataflow-between-stages.png b/docs/source/architecture/vllm-omni-dataflow-between-stages.png
index cdbc9a8b7b3..74abc81ff07 100644
Binary files a/docs/source/architecture/vllm-omni-dataflow-between-stages.png and b/docs/source/architecture/vllm-omni-dataflow-between-stages.png differ