This repository provides a collection of open-source Dify workflow examples, demonstrating practical applications of Dify's AI application DSL. It's designed for developers new to Dify, enterprise teams standardizing processes, educators, and the open-source community looking to collaborate on automation solutions. The primary benefit is enabling users to quickly build and deploy complex AI workflows by leveraging pre-built, commented examples.

How It Works

The project showcases Dify workflows that integrate with external services via Flask backends and HTTP requests. These workflows leverage Large Language Models (LLMs) for tasks like data generation, content extraction, summarization, and translation. Examples include generating Excel files, creating PPTs, producing Word documents, recognizing invoice data, analyzing GitHub projects, generating mind maps, and processing web content for knowledge bases. The approach emphasizes "workflow-as-code" for standardization and reusability.

Quick Start & Requirements

• Installation: Download specific workflow folders, navigate into their respective service directories (e.g., Excel_Flask_Dify/Excel_flask_Service), and run the Flask service using python3 <service_script.py>. Then, import the corresponding .yml DSL file into Dify and configure HTTP request nodes with the Flask service URL.
• Prerequisites: Python 3, Dify instance, Flask, and specific libraries for each workflow (e.g., openpyxl for Excel, python-docx for Word, Pillow for image processing). Some workflows may require specific LLM models or external tools like ComfyUI.
• Resources: Requires a running Dify instance and local execution of Flask services for each workflow. Setup time varies per workflow but involves downloading, configuring, and running a local service.

Highlighted Details

• Demonstrates practical AI applications: Excel generation, PPT creation, Word document generation, invoice OCR, GitHub project summarization, mind map generation, and more.
• Each workflow is accompanied by a Bilibili video tutorial by the author.
• Workflows are provided as .yml DSL files for easy import into Dify.
• Includes examples of integrating with external tools like ComfyUI for image generation.

Maintenance & Community

• Authored by BannyLon7, with contributions welcomed via Issues and Pull Requests.
• Contact information provided includes WeChat and a public WeChat account.

Licensing & Compatibility

• Licensed under the MIT License. This permissive license allows for commercial use, modification, and distribution without significant restrictions, making it highly compatible with closed-source projects.

Limitations & Caveats

• Workflows are provided as-is and may require adjustments due to Dify version upgrades or changes in LLM models.
• Some workflows rely on locally deployed services (e.g., Flask, ComfyUI), requiring users to set up and manage these dependencies.
• The effectiveness of certain workflows, like image generation, depends on the specific models and configurations used.