Does Azure OpenAI support PDF upload for GPT 4o?

Hi Zachary Xi

Azure OpenAI does not currently support direct PDF file input in the same way as OpenAI's API. However, there are workarounds using GPT-4o's vision capabilities to extract structured data from PDFs. There are alternative ways to process PDFs effectively using GPT-4o’s vision capabilities. Since native PDF handling is not yet available, users typically rely on two main approaches: text extraction and image-based analysis.
The text extraction method involves using tools such as Azure Document Intelligence or Python libraries like PyMuPDF and pdfplumber to extract readable text from PDFs, which can then be passed to Azure OpenAI for summarization or structured data extraction.
Meanwhile, the image-based approach converts PDF pages into images using libraries like pdf2image, enabling GPT-4o to analyze visual data and extract relevant information. Microsoft has provided an official sample demonstrating how GPT-4o can extract structured JSON data from PDFs, particularly for complex documents such as invoices.
Reference : Azure open ai

Thanks

Hi Zachary Xi

Just checking in to see if the below answer helped. If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.

Thanks

Hi Ravada, do you know when direct PDF input will be supported by Azure OpenAI? Can we expect it in the next few weeks/month? Thanks.

Hi LittleBear
Azure OpenAI provides a diverse suite of language models tailored for various applications, ensuring optimized performance across different tasks. The latest addition, GPT-4.1, represents a significant advancement in AI capabilities, supporting both text and image inputs while enabling chat completions, streaming responses, and structured outputs. With an extensive context window exceeding 1 million tokens and the ability to generate up to 32,768 output tokens, GPT-4.1 enhances long-context understanding and complex reasoning. For developers seeking faster response times, GPT-4.1-nano offers similar capabilities with accelerated processing speeds, making it ideal for real-time applications. Additionally, Azure OpenAI provides specialized models such as computer-use-preview, designed for automated system interactions, and audio-focused models capable of speech recognition and generation. These models are accessible via the Azure OpenAI API, which supports multiple interaction methods, including chat completions, streaming responses, and structured outputs, allowing seamless integration into applications using SDKs and API endpoints. You can explore more details on Microsoft Learn.
Thanks

Hi Zachary Xi

Following up to see if the above answer was helpful. If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.

Thanks