Help: Read PDF file into an Chatbot or Agent

[rmaddelein-hash]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing? Tell me about your story.

Hello,

I wish read a PDF file into an chatbot or agent, but this extension is not allowed.

Dify version: 1.11.1 (self-hosted)
Model (vision): Mistral large/GPT-4.1

• Is it normally, because this feature is not create?
• Is it a bug, because i use a bad model?

Thank you

2. Additional context or comments

No response

[dosubot[bot]]

Hi @rmaddelein-hash! I'm Dosu and I’m helping the dify team.

Reading PDF files into a chatbot or agent is supported in Dify v1.11.1. The platform uses a built-in PDF extractor to pull text from PDF files—this works for workflows, knowledge bases, and agents, and is not restricted by your model choice (Mistral large or GPT-4.1 are both fine) as long as the downstream workflow supports it. There is no need for a separate extension for basic text extraction from PDFs, and there is no bug preventing this feature for your models [source].

However, only text is extracted from PDFs. If your PDF contains images or scanned pages (not selectable text), Dify will not process or extract that content—no OCR or image extraction is performed. For image-based PDFs, you’ll need to use an external OCR tool to convert images to text before uploading to Dify [source].

To set this up, use the Document Extractor node in your workflow or agent pipeline. Upload your PDF, and Dify will extract the text for further processing by your LLM. If you need to handle large documents, consider chunking the extracted text before sending it to the model [source].

If you’re seeing an error like “extension not allowed,” double-check that the file input variable is configured to accept PDFs and that you’re not hitting a frontend validation issue. The backend supports PDF ingestion natively.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[rmaddelein-hash]

Hi,

The first source is for a RAG, used in knowledge base. My problem is to upload a PDF file with an agent.

With a model visio, i can't upload PDF file.

Is it normally, because this feature is not create?
Is it a bug, because i use a bad model?

[dosubot[bot]]

This is expected behavior—Dify agent applications do not currently support general file uploads like PDFs, even when using vision models such as Mistral large or GPT-4.1. Enabling vision mode only allows image uploads, not other file types. This is a product limitation, not a bug or a model compatibility issue. If you need to upload PDFs, you can use a Chatflow application (where file upload is supported), or build a custom frontend using Dify's backend APIs to handle file uploads programmatically. Multimodal/vision support for agent nodes is on the roadmap but not available yet as of late 2025 [source, source]. If you want guidance on alternative workflows or custom integrations, let me know!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}