diff --git a/uniflow/flow/config.py b/uniflow/flow/config.py
index 603d1877..cb9870c2 100644
--- a/uniflow/flow/config.py
+++ b/uniflow/flow/config.py
@@ -50,6 +50,15 @@ class ExtractPDFConfig(ExtractConfig):
     splitter: str = PARAGRAPH_SPLITTER
 
 
+@dataclass
+class ExtractS3PDFConfig(ExtractConfig):
+    """Extract S3 PDF Config Class."""
+
+    flow_name: str = "ExtractS3PDFFlow"
+    model_config: ModelConfig = field(default_factory=NougatModelConfig)
+    splitter: str = PARAGRAPH_SPLITTER
+
+
 @dataclass
 class ExtractImageConfig(ExtractConfig):
     """Extract Image Config Class"""
diff --git a/uniflow/flow/extract/extract_pdf_flow.py b/uniflow/flow/extract/extract_pdf_flow.py
index 4d213111..0ae2209d 100644
--- a/uniflow/flow/extract/extract_pdf_flow.py
+++ b/uniflow/flow/extract/extract_pdf_flow.py
@@ -5,6 +5,7 @@
 from uniflow.constants import EXTRACT
 from uniflow.flow.flow import Flow
 from uniflow.node import Node
+from uniflow.op.extract.load.aws.s3_op import ExtractS3PDFOp
 from uniflow.op.extract.load.pdf_op import ExtractPDFOp, ProcessPDFOp
 from uniflow.op.extract.split.constants import PARAGRAPH_SPLITTER
 from uniflow.op.extract.split.splitter_factory import SplitterOpsFactory
@@ -50,3 +51,28 @@ def run(self, nodes: Sequence[Node]) -> Sequence[Node]:
         nodes = self._process_pdf_op(nodes)
         nodes = self._split_op(nodes)
         return nodes
+
+
+class ExtractS3PDFFlow(ExtractPDFFlow):
+    """Extract S3 PDF Flow Class."""
+
+    def __init__(
+        self,
+        model_config: Dict[str, Any],
+        splitter: str = PARAGRAPH_SPLITTER,
+    ) -> None:
+        """Extract S3 PDF Flow Constructor.
+
+        Args:
+            model_config (Dict[str, Any]): Model config.
+            splitter (str): Splitter to use. Defaults to "".
+        """
+        super().__init__(model_config=model_config, splitter=splitter)
+        self._extract_pdf_op = ExtractS3PDFOp(
+            name="extract_s3_pdf_op",
+            model=LLMDataPreprocessor(
+                model_config=model_config,
+            ),
+        )
+        self._process_pdf_op = ProcessPDFOp(name="process_pdf_op")
+        self._split_op = SplitterOpsFactory.get(splitter)
diff --git a/uniflow/op/extract/load/aws/s3_op.py b/uniflow/op/extract/load/aws/s3_op.py
index 06c25ffd..256f04b1 100644
--- a/uniflow/op/extract/load/aws/s3_op.py
+++ b/uniflow/op/extract/load/aws/s3_op.py
@@ -6,6 +6,7 @@
 from typing import Sequence
 
 from uniflow.node import Node
+from uniflow.op.model.abs_llm_processor import AbsLLMProcessor
 from uniflow.op.op import Op
 
 logger = logging.getLogger(__name__)
@@ -66,3 +67,49 @@ def __call__(self, nodes: Sequence[Node]) -> Sequence[Node]:
                 )
             )
         return output_nodes
+
+
+class ExtractS3PDFOp(ExtractS3Op):
+    """Op to download and process PDF from s3."""
+
+    def __init__(self, model: AbsLLMProcessor, name: str = "extract_s3_pdf_op") -> None:
+        super().__init__(name=name)
+        self._model = model
+
+    def __call__(self, nodes: Sequence[Node]) -> Sequence[Node]:
+        """Run Model Op.
+
+        Args:
+            nodes (Sequence[Node]): Nodes to run.
+
+        Returns:
+            Sequence[Node]: Nodes after running.
+        """
+        output_nodes = []
+
+        for node in nodes:
+            value_dict = copy.deepcopy(node.value_dict)
+            # create local file path if not exists
+            if os.path.exists(self.LOCAL_FILE_PATH) is False:
+                os.makedirs(self.LOCAL_FILE_PATH)
+            filename = os.path.join(self.LOCAL_FILE_PATH, value_dict["key"])
+
+            logger.info("Downloading %s to %s", value_dict["key"], filename)
+            self._s3_client.download_file(
+                Bucket=value_dict["bucket"],
+                Key=value_dict["key"],
+                Filename=filename,
+            )
+
+            filename = {"pdf": filename}
+            value_dict = self._model.run(filename)
+            text = value_dict["response"][0]
+
+            output_nodes.append(
+                Node(
+                    name=self.unique_name(),
+                    value_dict={"text": text},
+                    prev_nodes=[node],
+                )
+            )
+        return output_nodes
diff --git a/uniflow/op/extract/load/pdf_op.py b/uniflow/op/extract/load/pdf_op.py
index 1ba85a3a..b6cff430 100644
--- a/uniflow/op/extract/load/pdf_op.py
+++ b/uniflow/op/extract/load/pdf_op.py
@@ -5,6 +5,7 @@
 from typing import Sequence
 
 from uniflow.node import Node
+from uniflow.op.extract.load.utils import read_file
 from uniflow.op.model.abs_llm_processor import AbsLLMProcessor
 from uniflow.op.op import Op
 
@@ -34,6 +35,7 @@ def __call__(self, nodes: Sequence[Node]) -> Sequence[Node]:
         output_nodes = []
         for node in nodes:
             value_dict = copy.deepcopy(node.value_dict)
+            value_dict = read_file(value_dict)
             value_dict = self._model.run(value_dict)
             text = value_dict["response"][0]
             output_nodes.append(