PDFToImageContent

Convert PDF files to ImageContent objects for multimodal AI processing. Each converted page becomes a separate ImageContent object containing base64-encoded image data and associated metadata.

PDFToImageContent reads PDF files and converts specified pages into images, creating ImageContent objects containing base64-encoded image data and metadata. Each converted page becomes a separate ImageContent object with metadata indicating the source file and page number.

Key Features

• Converts PDF pages to base64-encoded ImageContent objects for multimodal AI models.
• Supports specific page selection and page ranges.
• Optional image resizing to reduce file size while maintaining aspect ratio.
• Configurable detail level for optimization with OpenAI vision models.

Configuration

1. Drag the PDFToImageContent component onto the canvas from the Component Library.
2. Click on the component to open the configuration panel.
3. Configure the component settings:
   • Set the Page Range to specify which pages to convert. Accepts page numbers and ranges, for example ['1-3', '5', '8']. If not set, all pages are converted.
   • Set the Detail level for images (auto, high, or low). This is passed to the created ImageContent objects and is only supported by OpenAI.
   • Set the Size to resize images to the specified dimensions (width, height) while maintaining aspect ratio. This reduces file size, memory usage, and processing time.

Connections

PDFToImageContent accepts a list of file paths or ByteStream objects through its sources input. It outputs a list of ImageContent objects, one per converted page.

It typically connects with:

• FilesInput or FileTypeRouter: receives PDF files.
• ChatPromptBuilder: sends extracted page images to include in multimodal prompts.

Source Code

To check this component's source code, open pdf_to_image.py in the Haystack repository.

Usage Examples

Basic Configuration

  PDFToImageContent:
    type: haystack.components.converters.image.PDFToImageContent
    init_parameters:
      detail: auto
      page_range:
      - 1-3

Using the Component in a Pipeline

This example shows a query pipeline that uses PDFToImageContent to convert uploaded PDF pages to images for multimodal AI processing. The pipeline extracts specified pages from PDF documents, converts them to ImageContent objects, and sends them to a vision-enabled chat model for analysis.

# haystack-pipeline
components:
  PDFToImageContent:
    type: haystack.components.converters.image.PDFToImageContent
    init_parameters:
      detail: auto
      size:
      page_range:
        - "1-3"
  ChatPromptBuilder:
    type: haystack.components.builders.chat_prompt_builder.ChatPromptBuilder
    init_parameters:
      template:
        - role: system
          content: >-
            You are an AI assistant that analyzes PDF documents. Examine the
            provided PDF pages and answer questions about their content,
            layout, text, images, and any other visible elements.
        - role: user
          content: "{{ question }}"
          images: "{{ images }}"
      required_variables:
        - question
        - images
  OpenAIChatGenerator:
    type: haystack.components.generators.chat.openai.OpenAIChatGenerator
    init_parameters:
      api_key:
        type: env_var
        env_vars:
          - OPENAI_API_KEY
        strict: true
      model: gpt-4o
      generation_kwargs:
  OutputAdapter:
    type: haystack.components.converters.output_adapter.OutputAdapter
    init_parameters:
      template: "{{ replies[0].text }}"
      output_type: str
  DeepsetAnswerBuilder:
    type: deepset_cloud_custom_nodes.augmenters.deepset_answer_builder.DeepsetAnswerBuilder
    init_parameters:
      reference_pattern:

connections:
  - sender: PDFToImageContent.image_contents
    receiver: ChatPromptBuilder.images
  - sender: ChatPromptBuilder.prompt
    receiver: OpenAIChatGenerator.messages
  - sender: OpenAIChatGenerator.replies
    receiver: OutputAdapter.replies
  - sender: OutputAdapter.output
    receiver: DeepsetAnswerBuilder.replies

inputs:
  query:
    - ChatPromptBuilder.question
  files:
    - PDFToImageContent.sources

outputs:
  answers: DeepsetAnswerBuilder.answers

Parameters

Inputs

Parameter	Type	Default	Description
sources	List[Union[str, Path, ByteStream]]		List of PDF file paths or ByteStream objects to convert.
meta	Optional[Union[Dict[str, Any], List[Dict[str, Any]]]]	None	Optional metadata to attach to the ImageContent objects. This value can be a list of dictionaries or a single dictionary. If it's a single dictionary, its content is added to the metadata of all produced ImageContent objects. If it's a list, its length must match the number of sources as they're zipped together. For ByteStream objects, their meta is added to the output ImageContent objects.
detail	Optional[Literal["auto", "high", "low"]]	None	Optional detail level of the image (only supported by OpenAI). This is passed to the created ImageContent objects. If not provided, the detail level is the one set in the constructor.
size	Optional[Tuple[int, int]]	None	If provided, resizes the image to fit within the specified dimensions (width, height) while maintaining aspect ratio. If not provided, the size value is the one set in the constructor.
page_range	Optional[List[Union[str, int]]]	None	List of page numbers and page ranges to convert to images. If not provided, the page range is the one set in the pipeline configuration.

Outputs

Parameter	Type	Description
image_contents	List[ImageContent]	A list of ImageContent objects created from the PDF pages.

Init Parameters

These are the parameters you can configure in Pipeline Builder:

Parameter	Type	Default	Description
detail	Optional[Literal["auto", "high", "low"]]	None	Optional detail level of the image (only supported by OpenAI). Possible values: "auto", "high", or "low". This is passed to the created ImageContent objects.
size	Optional[Tuple[int, int]]	None	If provided, resizes the image to fit within the specified dimensions (width, height) while maintaining aspect ratio. This reduces file size, memory usage, and processing time.
page_range	Optional[List[Union[str, int]]]	None	List of page numbers and page ranges to convert to images. Page numbers start at 1. If None, all pages in the PDF are converted. Also accepts printable range strings, for example: ['1-3', '5', '8', '10-12'].

Run Method Parameters

These are the parameters you can configure for the component's run() method. This means you can pass these parameters at query time through the API, in Playground, or when running a job. For details, see Modify Pipeline Parameters at Query Time.

Parameter	Type	Default	Description
sources	List[Union[str, Path, ByteStream]]		List of PDF file paths or ByteStream objects to convert.
meta	Optional[Union[Dict[str, Any], List[Dict[str, Any]]]]	None	Optional metadata to attach to the ImageContent objects. This value can be a list of dictionaries or a single dictionary. If it's a single dictionary, its content is added to the metadata of all produced ImageContent objects. If it's a list, its length must match the number of sources as they're zipped together. For ByteStream objects, their meta is added to the output ImageContent objects.
detail	Optional[Literal["auto", "high", "low"]]	None	Optional detail level of the image (only supported by OpenAI). This is passed to the created ImageContent objects. If not provided, the detail level is the one set in the constructor.
size	Optional[Tuple[int, int]]	None	If provided, resizes the image to fit within the specified dimensions (width, height) while maintaining aspect ratio. If not provided, the size value is the one set in the constructor.
page_range	Optional[List[Union[str, int]]]	None	List of page numbers and page ranges to convert to images. Page numbers start at 1. If None, all pages in the PDF are converted. Pages outside the valid range are skipped with a warning. Also accepts printable range strings, for example: ['1-3', '5', '8', '10-12'].