Essential skills for Dify users: Full steps for low-cost integration of third - party APIs

Issuing time:2026-01-02 19:00


💡Foreword: While Dify's built-in official API is convenient, its cost can be quite high with frequent, long-term use! Today, I'm sharing a money-saving tip – use the Nebula API. It's only 1/5 or even less of the official API, yet the performance is almost identical. Frequent users can save enough to buy bubble tea!




🔍 First, let's understand: What is Dify?



In short: Dify is a "visual + plug-in" AI application building platform. Even those with zero coding experience can easily use large models to quickly build their own AI tools (such as chatbots, text processing tools, etc.). Even beginners can get started!




🛠️ ​​Part Two: Step-by-Step Guide to Integrating Third-Party Models (Super Easy!)



Step 1: Prepare a third-party API account


First, register a third-party API service provider account.

👉 https://openai-nebula.com

(Subsequent operations will follow this example; the registration process is super simple!)


Step 2: Enter Dify Model Supplier Settings

  1. Open the Dify platform, log in, and find the "Settings" page;

  2. Select "Model Supplier" in the left-hand menu to enter the plugin installation interface.


Step 3: Install compatibility plugins

  1. Scroll down the page and find the "OpenAI-API-compatible" plugin;

  2. This plugin is incredibly useful! It supports various functions such as dialogue, text embedding, and speech-to-text conversion, allowing Dify to seamlessly integrate with large third-party models. Simply click "Install"!


Step 4: Configure key parameters (Important!)

After installation, click "Add Model". You will need to... https://openai-nebula.com To obtain 3 core parameters, let's go step by step:


1. Obtain the API Key (📝 just copy and paste)

  • Open api.chatterai.top and click "Tokens" in the top navigation;

  • Copy the "token" generated on the page and paste it into the "API Key" field of the Dify parameters.


2. Retrieve the model name (warning:️ The two fields must match)

  • Back https://openai-nebula.comGo to the "Account" page;

  • In the blue "Model List" at the bottom of the page, select the model you need (such as Gemini, GPT series), and click to copy the name;

  • Enter the copied name into Dify's "Model Name" and "Model Name in API endpoint" respectively (make sure you enter them correctly!).



3. Enter the API address (Just copy this!)


4. Save configuration

After filling in all the parameters, click "Save"—don't go! The third-party model has been successfully integrated into Dify!




📊III. Actual Test: Surprising Results & Cost!



Test steps:

  1. Open the workflow task you created in Dify;

  2. Replace the existing model with the newly integrated third-party model;

  3. Click "Start Run" and wait for the execution result.

Follow the steps and you can get low-cost, large-scale model integration done in just a few minutes!


Surprising result:

✅ Call successful! Task executed smoothly, response speed is identical to the official version;

💰 The cost is outrageous! Check the api.chatterai.top backend: calling the latest Gemini model consumed over 80,000 tokens, but only cost 0.1 yuan! (Compared to the official price, this is a saving of over 80%, incredibly affordable for frequent use!)



NEBULA DATA, headquartered in Singapore with branches in Jakarta, Guangzhou, Shanghai, and Hong Kong, is a leading provider of integrated cloud and network solutions and intelligent computing power, focusing on serving Chinese enterprises expanding globally. The company has independently developed the Nebula API unified LLM interface platform, aggregating mainstream global models such as Claude and Gemini, supporting one-click access, and the Nebula Lab one-stop AI content generation and marketing platform, widely used in the e-commerce sector. It also provides integrated solutions for global AIDC (AI Computing Center) + low-latency networks + converged cloud management, empowering enterprises to embrace AI and expand globally.