Dify
Dify is an open-source LLM application development and orchestration platform (40k+ Stars). Combining the concepts of BaaS and LLMOps, it supports visual orchestration of autonomous Agents, enterprise knowledge bases (RAG), complex Workflows, and Tool Calling. Integrate with OpenRely API to aggregate top global foundation model computing power in one stop.
🔑 Prerequisites
sk-...🚀 Steps
Log in to Dify Platform
Open Dify Cloud in your browser or log in to your self-hosted Dify Web console.
Go to Model Provider Settings
- Click your profile avatar in the top-right corner of the Dify page;
- Select Settings from the dropdown menu;
- Click Model Providers in the left navigation bar.
Add OpenRely Custom Model Provider
In the model provider list:
- Locate the OpenAI-API-compatible card and click Add Model;
- Fill in the parameters according to the table below:
| Configuration Item | Recommended Value | Description |
|---|---|---|
| Model Type | LLM | Large Language Model |
| Model Name | claude or gpt | Enter any model identifier you need to invoke (e.g., claude, gpt, etc.) |
| API Base URL | https://api.openrely.ai/v1 | Must include the /v1 suffix |
| API Key | sk-... | Dedicated access Token created in the OpenRely console |
- Click the Save button in the bottom-right corner, and Dify will automatically run a connectivity test.
Invoke Models in Workflows and Applications
Once configured, you can use the model in any scenario within Dify:
- Studio Application Orchestration: Create Chat Apps or Agents, and select the OpenRely model from the top model dropdown;
- Workflow Orchestration: Freely select the model in LLM nodes to leverage high-concurrency, ultra-fast model inference capabilities.
💡 Core Features
| Feature Module | Description |
|---|---|
| Visual Workflow | Chain Prompts, conditional branches, code execution nodes, and external HTTP requests like building blocks |
| Enterprise Knowledge Base (RAG) | Supports batch importing PDF, Word, and Notion documents, with text chunking, vector retrieval, and semantic reranking |
| Tool Ecosystem & Plugins | Native support for Google Search, code sandboxes, DALL-E image generation, and custom OpenAPI tool integration |
| Full API Access & Embedding | Orchestrated Agents support one-click generation of standard RESTful APIs or Web widgets for third-party system integration |
❓ FAQ
Encountering other API errors or billing questions? Please refer to the complete FAQ and Troubleshooting Guide.
Was this page helpful?
LunaTranslator
LunaTranslator is a free, open-source translation tool for visual novels (GalGames) and multimedia. It supports text hooking, offline OCR, and LLM translation engines, connecting with OpenRely API for natural and fluent native-level translation.
Memoh
Memoh is an open-source, self-hosted AI agent platform. Each bot runs in an isolated container with persistent memory and an independent file system, supporting integration with 9 IM channels including Telegram, Discord, Feishu, and WeChat, and connects to the OpenRely API for powerful LLM capabilities.