docling-project · maxdswain · Mar 22, 2026 · Mar 22, 2026 · Mar 22, 2026
diff --git a/docling/cli/main.py b/docling/cli/main.py
@@ -428,6 +428,12 @@ def convert(  # noqa: C901
             ..., help="If enabled, the bitmap content will be processed using OCR."
         ),
     ] = True,
+    layout: Annotated[
+        bool,
+        typer.Option(
+            ..., help=("If enabled document layout analysis processing will be done.")
+        ),
+    ] = True,
     force_ocr: Annotated[
         bool,
         typer.Option(
@@ -733,6 +739,7 @@ def convert(  # noqa: C901
                 enable_remote_services=enable_remote_services,
                 accelerator_options=accelerator_options,
                 do_ocr=ocr,
+                do_layout=layout,
                 ocr_options=ocr_options,
                 do_table_structure=tables,
                 do_code_enrichment=enrich_code,

diff --git a/docling/datamodel/pipeline_options.py b/docling/datamodel/pipeline_options.py
@@ -1416,6 +1416,15 @@ class PdfPipelineOptions(PaginatedPipelineOptions):
             )
         ),
     ] = True
+    do_layout: Annotated[
+        bool,
+        Field(
+            description=(
+                "Enable document layout analysis to detect and classify page regions such as text blocks, headings, figures, "
+                "tables, and other structural elements. Required for accurate content segmentation and reading-order reconstruction."
+            )
+        ),
+    ] = True
     do_code_enrichment: Annotated[
         bool,
         Field(

diff --git a/docling/experimental/datamodel/threaded_layout_vlm_pipeline_options.py b/docling/experimental/datamodel/threaded_layout_vlm_pipeline_options.py
@@ -23,6 +23,7 @@ class ThreadedLayoutVlmPipelineOptions(PaginatedPipelineOptions):
     vlm_options: Union[InlineVlmOptions, ApiVlmOptions] = GRANITEDOCLING_TRANSFORMERS
 
     # Layout model configuration
+    do_layout: bool = True
     layout_options: LayoutOptions = LayoutOptions(
         model_spec=DOCLING_LAYOUT_HERON, skip_cell_assignment=True
     )

diff --git a/docling/experimental/models/table_crops_layout_model.py b/docling/experimental/models/table_crops_layout_model.py
@@ -33,10 +33,12 @@ def __init__(
         artifacts_path: Optional[Path],
         accelerator_options: AcceleratorOptions,
         options: TableCropsLayoutOptions,
+        enabled: bool,
         enable_remote_services: bool = False,
     ):
         _ = enable_remote_services
         self.options = options
+        self.enabled = enabled
         self.artifacts_path = artifacts_path
         self.accelerator_options = accelerator_options
 

diff --git a/docling/experimental/pipeline/threaded_layout_vlm_pipeline.py b/docling/experimental/pipeline/threaded_layout_vlm_pipeline.py
@@ -75,6 +75,7 @@ def _init_models(self) -> None:
             artifacts_path=art_path,
             accelerator_options=self.pipeline_options.accelerator_options,
             options=self.pipeline_options.layout_options,
+            enabled=self.pipeline_options.do_layout,
         )
 
         # VLM model based on options type

diff --git a/docling/models/base_layout_model.py b/docling/models/base_layout_model.py
@@ -13,6 +13,8 @@
 class BaseLayoutModel(BasePageModel, BaseModelWithOptions, ABC):
     """Shared interface for layout models."""
 
+    enabled: bool
+
     @classmethod
     @abstractmethod
     def get_options_type(cls) -> Type[BaseLayoutOptions]:
@@ -31,6 +33,10 @@ def __call__(
         conv_res: ConversionResult,
         page_batch: Iterable[Page],
     ) -> Iterable[Page]:
+        if not self.enabled:
+            yield from page_batch
+            return
+
         pages = list(page_batch)
         predictions = self.predict_layout(conv_res, pages)
 

diff --git a/docling/models/stages/layout/layout_model.py b/docling/models/stages/layout/layout_model.py
@@ -51,12 +51,14 @@ def __init__(
         artifacts_path: Optional[Path],
         accelerator_options: AcceleratorOptions,
         options: LayoutOptions,
+        enabled: bool,
         enable_remote_services: bool = False,
     ):
         from docling_ibm_models.layoutmodel.layout_predictor import LayoutPredictor
 
         _ = enable_remote_services
         self.options = options
+        self.enabled = enabled
 
         device = decide_device(accelerator_options.device)
         layout_model_config = options.model_spec

diff --git a/docling/models/stages/layout/layout_object_detection_model.py b/docling/models/stages/layout/layout_object_detection_model.py
@@ -35,9 +35,11 @@ def __init__(
         artifacts_path: Optional[Path],
         accelerator_options: AcceleratorOptions,
         options: LayoutObjectDetectionOptions,
+        enabled: bool,
         enable_remote_services: bool = False,
     ) -> None:
         self.options = options
+        self.enabled = enabled
 
         self.engine: BaseObjectDetectionEngine = create_object_detection_engine(
             options=options.engine_options,

diff --git a/docling/models/stages/page_assemble/page_assemble_model.py b/docling/models/stages/page_assemble/page_assemble_model.py
@@ -43,8 +43,9 @@ class PageAssembleOptions(BaseModel):
 
 
 class PageAssembleModel(BasePageModel):
-    def __init__(self, options: PageAssembleOptions):
+    def __init__(self, options: PageAssembleOptions, enabled: bool):
         self.options = options
+        self.enabled = enabled
 
     def sanitize_text(self, lines):
         if len(lines) == 0:
@@ -88,6 +89,10 @@ def sanitize_text(self, lines):
     def __call__(
         self, conv_res: ConversionResult, page_batch: Iterable[Page]
     ) -> Iterable[Page]:
+        if not self.enabled:
+            yield from page_batch
+            return
+
         for page in page_batch:
             assert page._backend is not None
             if not page._backend.is_valid():

diff --git a/docling/pipeline/legacy_standard_pdf_pipeline.py b/docling/pipeline/legacy_standard_pdf_pipeline.py
@@ -64,6 +64,7 @@ def __init__(self, pipeline_options: PdfPipelineOptions):
         )
         layout_model = layout_factory.create_instance(
             options=pipeline_options.layout_options,
+            enabled=pipeline_options.do_layout,
             artifacts_path=self.artifacts_path,
             accelerator_options=pipeline_options.accelerator_options,
             enable_remote_services=pipeline_options.enable_remote_services,
@@ -93,7 +94,9 @@ def __init__(self, pipeline_options: PdfPipelineOptions):
             # Table structure model
             table_model,
             # Page assemble
-            PageAssembleModel(options=PageAssembleOptions()),
+            PageAssembleModel(
+                options=PageAssembleOptions(), enabled=pipeline_options.do_layout
+            ),
         ]
 
         self.enrichment_pipe = [

diff --git a/docling/pipeline/standard_pdf_pipeline.py b/docling/pipeline/standard_pdf_pipeline.py
@@ -488,6 +488,7 @@ def _init_models(self) -> None:
         )
         self.layout_model = layout_factory.create_instance(
             options=self.pipeline_options.layout_options,
+            enabled=self.pipeline_options.do_layout,
             artifacts_path=art_path,
             accelerator_options=self.pipeline_options.accelerator_options,
             enable_remote_services=self.pipeline_options.enable_remote_services,
@@ -502,7 +503,9 @@ def _init_models(self) -> None:
             accelerator_options=self.pipeline_options.accelerator_options,
             enable_remote_services=self.pipeline_options.enable_remote_services,
         )
-        self.assemble_model = PageAssembleModel(options=PageAssembleOptions())
+        self.assemble_model = PageAssembleModel(
+            options=PageAssembleOptions(), enabled=self.pipeline_options.do_layout
+        )
         self.reading_order_model = ReadingOrderModel(options=ReadingOrderOptions())
 
         # --- optional enrichment ------------------------------------------------

diff --git a/docs/examples/custom_convert.py b/docs/examples/custom_convert.py
@@ -76,6 +76,22 @@ def main():
     #     }
     # )
 
+    # PyPdfium without EasyOCR and Layout Processing
+    # --------------------
+    # pipeline_options = PdfPipelineOptions()
+    # pipeline_options.do_ocr = False
+    # pipeline_options.do_layout = False
+    # pipeline_options.do_table_structure = True
+    # pipeline_options.table_structure_options = TableStructureOptions(do_cell_matching=False)
+
+    # doc_converter = DocumentConverter(
+    #     format_options={
+    #         InputFormat.PDF: PdfFormatOption(
+    #             pipeline_options=pipeline_options, backend=PyPdfiumDocumentBackend
+    #         )
+    #     }
+    # )
+
     # PyPdfium with EasyOCR
     # -----------------
     # pipeline_options = PdfPipelineOptions()
@@ -104,6 +120,20 @@ def main():
     #     }
     # )
 
+    # Docling Parse without EasyOCR and Layout Processing
+    # -------------------------
+    # pipeline_options = PdfPipelineOptions()
+    # pipeline_options.do_ocr = False
+    # pipeline_options.do_layout = False
+    # pipeline_options.do_table_structure = True
+    # pipeline_options.table_structure_options = TableStructureOptions(do_cell_matching=True)
+
+    # doc_converter = DocumentConverter(
+    #     format_options={
+    #         InputFormat.PDF: PdfFormatOption(pipeline_options=pipeline_options)
+    #     }
+    # )
+
     # Docling Parse with EasyOCR (default)
     # -------------------------------
     # Enables OCR and table structure with EasyOCR, using automatic device

diff --git a/tests/test_page_assemble_model.py b/tests/test_page_assemble_model.py
@@ -10,7 +10,7 @@
 
 @pytest.fixture
 def model() -> PageAssembleModel:
-    return PageAssembleModel(options=PageAssembleOptions())
+    return PageAssembleModel(options=PageAssembleOptions(), enabled=True)
 
 
 class TestSanitizeTextLigatures: