Integrate NVIDIA Ingest with Langflow

The NVIDIA Ingest component integrates with the NVIDIA nv-ingest microservice for data ingestion, processing, and extraction of text files.

The nv-ingest service supports multiple extraction methods for PDF, DOCX, and PPTX file types, and includes pre- and post-processing services like splitting, chunking, and embedding generation.

The NVIDIA Ingest component imports the NVIDIA Ingestor client, ingests files with requests to the NVIDIA ingest endpoint, and outputs the processed content as a list of Data objects. Ingestor accepts additional configuration options for data extraction from other text formats. To configure these options, see the component parameters.

Prerequisites

• An NVIDIA Ingest endpoint. For more information on setting up an NVIDIA Ingest endpoint, see the NVIDIA Ingest quickstart.

• The NVIDIA Ingest component requires the installation of additional dependencies to your Langflow environment. To install the dependencies in a virtual environment, run the following commands.

  • If you have the Langflow repository cloned and installed from source:
  
  

  _10

  source **YOUR_LANGFLOW_VENV**/bin/activate

  _10

  uv sync --extra nv-ingest

  _10

  uv run langflow run

  
  
  • If you are installing Langflow from the Python Package Index:
  
  

  _10

  source **YOUR_LANGFLOW_VENV**/bin/activate

  _10

  uv pip install --prerelease=allow 'langflow[nv-ingest]'

  _10

  uv run langflow run

  
  

Use the NVIDIA Ingest component in a flow

The NVIDIA Ingest component accepts Message inputs and outputs Data. The component calls a NVIDIA Ingest microservice's endpoint to ingest a local file and extract the text.

To use the NVIDIA Ingest component in your flow, follow these steps:

1. In the component library, click the NVIDIA Ingest component, and then drag it onto the canvas.
2. In the NVIDIA Ingestion URL field, enter the URL of the NVIDIA Ingest endpoint. Optionally, add the endpoint URL as a Global variable:
   1. Click Settings, and then click Global Variables.
   2. Click Add New.
   3. Name your variable. Paste your endpoint in the Value field.
   4. In the Apply To Fields field, select the field you want to globally apply this variable to. In this case, select NVIDIA Ingestion URL.
   5. Click Save Variable.
3. In the Path field, enter the path to the file you want to ingest.
4. Select which text type to extract from the file. The component supports text, charts, and tables.
5. Select whether to split the text into chunks. Modify the splitting parameters in the component's Configuration tab.
6. Click Run to ingest the file.
7. To confirm the component is ingesting the file, open the Logs pane to view the output of the flow.
8. To store the processed data in a vector database, add an AstraDB Vector component to your flow, and connect the NVIDIA Ingest component to the AstraDB Vector component with a Data output.

10. Run the flow. Inspect your Astra DB vector database to view the processed data.

NVIDIA Ingest component parameters

The NVIDIA Ingest component has the following parameters.

For more information, see the NV-Ingest documentation.

Inputs

Name	Display Name	Info
base_url	NVIDIA Ingestion URL	The URL of the NVIDIA Ingestion API.
path	Path	File path to process.
extract_text	Extract Text	Extract text from documents. Default: True.
extract_charts	Extract Charts	Extract text from charts. Default: False.
extract_tables	Extract Tables	Extract text from tables. Default: True.
text_depth	Text Depth	The level at which text is extracted. Support for 'block', 'line', and 'span' varies by document type. Default: document.
split_text	Split Text	Split text into smaller chunks. Default: True.
split_by	Split By	How to split into chunks. 'size' splits by number of characters. Default: word.
split_length	Split Length	The size of each chunk based on the 'split_by' method. Default: 200.
split_overlap	Split Overlap	The number of segments to overlap from the previous chunk. Default: 20.
max_character_length	Max Character Length	The maximum number of characters in each chunk. Default: 1000.
sentence_window_size	Sentence Window Size	The number of sentences to include from previous and following chunks when split_by=sentence. Default: 0.

Outputs

The NVIDIA Ingest component outputs a list of Data objects where each object contains:

• text: The extracted content.
  • For text documents: The extracted text content.
  • For tables and charts: The extracted table/chart content.
• file_path: The source file name and path.
• document_type: The type of the document ("text" or "structured").
• description: Additional description of the content.

The output varies based on the document_type:

• Documents with document_type: "text" contain:

  • Raw text content extracted from documents, for example, paragraphs from PDFs or DOCX files.
  • Content stored directly in the text field.
  • Content extracted using the extract_text parameter.

• Documents with document_type: "structured" contain:

  • Text extracted from tables and charts and processed to preserve structural information.
  • Content extracted using the extract_tables and extract_charts parameters.
  • Content stored in the text field after being processed from the table_content metadata.

note

Images are currently not supported and will be skipped during processing.