diff --git a/pilot/embedding_engine/markdown_embedding.py b/pilot/embedding_engine/markdown_embedding.py
index e9a97dce9..2bbd20878 100644
--- a/pilot/embedding_engine/markdown_embedding.py
+++ b/pilot/embedding_engine/markdown_embedding.py
@@ -6,7 +6,7 @@ from typing import List
 import markdown
 from bs4 import BeautifulSoup
 from langchain.schema import Document
-from langchain.text_splitter import SpacyTextSplitter
+from langchain.text_splitter import SpacyTextSplitter, CharacterTextSplitter
 
 from pilot.configs.config import Config
 from pilot.embedding_engine import SourceEmbedding, register
@@ -30,12 +30,20 @@ class MarkdownEmbedding(SourceEmbedding):
     def read(self):
         """Load from markdown path."""
         loader = EncodeTextLoader(self.file_path)
-        textsplitter = SpacyTextSplitter(
-            pipeline="zh_core_web_sm",
-            chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
-            chunk_overlap=100,
-        )
-        return loader.load_and_split(textsplitter)
+        # text_splitter = SpacyTextSplitter(
+        #     pipeline="zh_core_web_sm",
+        #     chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+        #     chunk_overlap=100,
+        # )
+        if CFG.LANGUAGE == "en":
+            text_splitter = CharacterTextSplitter(
+                chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+                chunk_overlap=20,
+                length_function=len,
+            )
+        else:
+            text_splitter = CHNDocumentSplitter(pdf=True, sentence_size=1000)
+        return loader.load_and_split(text_splitter)
 
     @register
     def data_process(self, documents: List[Document]):
diff --git a/pilot/embedding_engine/pdf_embedding.py b/pilot/embedding_engine/pdf_embedding.py
index ea4276460..a51eccbda 100644
--- a/pilot/embedding_engine/pdf_embedding.py
+++ b/pilot/embedding_engine/pdf_embedding.py
@@ -4,10 +4,11 @@ from typing import List
 
 from langchain.document_loaders import PyPDFLoader
 from langchain.schema import Document
-from langchain.text_splitter import SpacyTextSplitter
+from langchain.text_splitter import SpacyTextSplitter, CharacterTextSplitter
 
 from pilot.configs.config import Config
 from pilot.embedding_engine import SourceEmbedding, register
+from pilot.embedding_engine.chn_document_splitter import CHNDocumentSplitter
 
 CFG = Config()
 
@@ -28,12 +29,20 @@ class PDFEmbedding(SourceEmbedding):
         # textsplitter = CHNDocumentSplitter(
         #     pdf=True, sentence_size=CFG.KNOWLEDGE_CHUNK_SIZE
         # )
-        textsplitter = SpacyTextSplitter(
-            pipeline="zh_core_web_sm",
-            chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
-            chunk_overlap=100,
-        )
-        return loader.load_and_split(textsplitter)
+        # textsplitter = SpacyTextSplitter(
+        #     pipeline="zh_core_web_sm",
+        #     chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+        #     chunk_overlap=100,
+        # )
+        if CFG.LANGUAGE == "en":
+            text_splitter = CharacterTextSplitter(
+                chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+                chunk_overlap=20,
+                length_function=len,
+            )
+        else:
+            text_splitter = CHNDocumentSplitter(pdf=True, sentence_size=1000)
+        return loader.load_and_split(text_splitter)
 
     @register
     def data_process(self, documents: List[Document]):
diff --git a/pilot/embedding_engine/ppt_embedding.py b/pilot/embedding_engine/ppt_embedding.py
index 485083d1c..4ff06c6b7 100644
--- a/pilot/embedding_engine/ppt_embedding.py
+++ b/pilot/embedding_engine/ppt_embedding.py
@@ -4,10 +4,11 @@ from typing import List
 
 from langchain.document_loaders import UnstructuredPowerPointLoader
 from langchain.schema import Document
-from langchain.text_splitter import SpacyTextSplitter
+from langchain.text_splitter import SpacyTextSplitter, CharacterTextSplitter
 
 from pilot.configs.config import Config
 from pilot.embedding_engine import SourceEmbedding, register
+from pilot.embedding_engine.chn_document_splitter import CHNDocumentSplitter
 
 CFG = Config()
 
@@ -25,12 +26,20 @@ class PPTEmbedding(SourceEmbedding):
     def read(self):
         """Load from ppt path."""
         loader = UnstructuredPowerPointLoader(self.file_path)
-        textsplitter = SpacyTextSplitter(
-            pipeline="zh_core_web_sm",
-            chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
-            chunk_overlap=200,
-        )
-        return loader.load_and_split(textsplitter)
+        # textsplitter = SpacyTextSplitter(
+        #     pipeline="zh_core_web_sm",
+        #     chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+        #     chunk_overlap=200,
+        # )
+        if CFG.LANGUAGE == "en":
+            text_splitter = CharacterTextSplitter(
+                chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+                chunk_overlap=20,
+                length_function=len,
+            )
+        else:
+            text_splitter = CHNDocumentSplitter(pdf=True, sentence_size=1000)
+        return loader.load_and_split(text_splitter)
 
     @register
     def data_process(self, documents: List[Document]):
diff --git a/pilot/embedding_engine/word_embedding.py b/pilot/embedding_engine/word_embedding.py
index 34fc48450..9668700a1 100644
--- a/pilot/embedding_engine/word_embedding.py
+++ b/pilot/embedding_engine/word_embedding.py
@@ -4,6 +4,7 @@ from typing import List
 
 from langchain.document_loaders import PyPDFLoader, UnstructuredWordDocumentLoader
 from langchain.schema import Document
+from langchain.text_splitter import CharacterTextSplitter
 
 from pilot.configs.config import Config
 from pilot.embedding_engine import SourceEmbedding, register
@@ -25,10 +26,15 @@ class WordEmbedding(SourceEmbedding):
     def read(self):
         """Load from word path."""
         loader = UnstructuredWordDocumentLoader(self.file_path)
-        textsplitter = CHNDocumentSplitter(
-            pdf=True, sentence_size=CFG.KNOWLEDGE_CHUNK_SIZE
-        )
-        return loader.load_and_split(textsplitter)
+        if CFG.LANGUAGE == "en":
+            text_splitter = CharacterTextSplitter(
+                chunk_size=CFG.KNOWLEDGE_CHUNK_SIZE,
+                chunk_overlap=20,
+                length_function=len,
+            )
+        else:
+            text_splitter = CHNDocumentSplitter(pdf=True, sentence_size=1000)
+        return loader.load_and_split(text_splitter)
 
     @register
     def data_process(self, documents: List[Document]):