Skip to content
4ALL API Docs
Main Site News Console

Dify

Dify is an open-source LLM application orchestration platform that integrates 4ALL API through the “OpenAI-API-compatible” provider.

Configuration Steps

  1. Go to Settings → Model Providers;
  2. Find OpenAI-API-compatible, then click Add Model:
    • Model Type: LLM (or Text Embedding, etc., choose as needed)
    • Model Name: such as gpt-5.6
    • API Key: your sk- token
    • API endpoint URL: https://api.4allapi.com/v1
  3. After saving, you can select this model from the model dropdown in the application orchestration;
  4. If you need multiple models, repeat the steps above to add them one by one.

Recommendations

  • The embedding model used for knowledge base retrieval can also be added from 4ALL API (choose Text Embedding for the type), and it can share the same token as the chat model for unified billing;
  • In workflows, different nodes can be bound to different models, allocated according to task complexity to control overall costs;
  • If self-hosted Dify is running on an internal network, make sure the server can access port 443 of api.4allapi.com.