Skip to main content
Dify configures providers at workspace level. After adding APIONE, select its model explicitly in the application or LLM node.

Add a model

  1. Open Settings → Model Provider in your workspace.
  2. Find OpenAI-API-compatible. If necessary, install it through Dify’s official plugin marketplace.
  3. Click Add Model and choose LLM.
  4. Enter an APIONE model ID supporting Chat Completions.
  5. Enter the APIONE key and set API Endpoint / Base URL to https://api.apione.ai/v1.
  6. Set context and output limits according to the model, then save.
If a field explicitly requires a complete endpoint, use https://api.apione.ai/v1/chat/completions and avoid duplicated paths.
Add an OpenAI-compatible APIONE model in Dify; use the actual model ID and limits.

Add an OpenAI-compatible APIONE model in Dify; use the actual model ID and limits.

Verify an application

Create a Chatbot or open an LLM workflow node and select the new model. Run a minimal text-only flow before adding knowledge bases, tools, or multimodal input.

Common problems

  • Validation fails on save: check outbound access from the Dify backend or plugin runtime, the key, and the model ID.
  • An app still uses the old model: select the model again in the LLM node and save the app.
  • Knowledge retrieval fails: embeddings, reranking, and chat are separate capabilities requiring separate configuration.
  • Plugin or self-hosted differences: use the field definitions for your Dify and provider-plugin versions.
Integration directory · Choose a model · Troubleshooting