openai-assistant

• Node.js (^20.0.0 version)
• npm (^10.0.0 version)
• NestJS (^10.0.0 version)
• OpenAI (^4.51.0 version)
• OpenAI API key

Open or create your NestJS application where you would like to integrate the AI Assistant. To create a new NestJS application, use the following command:

nest new project-name

Now you have to install the packages. Go to the next step.

Make sure you are in the root directory of your project. Install the library and openai package using npm:

npm i @boldare/openai-assistant openai --save

The library is installed but we have to configure it. Go to the next step.

Set up your environment variables, create environment variables in the .env file in the root directory of the project, and populate it with the necessary secrets. The assistant ID is optional and serves as a unique identifier for your assistant. When the environment variable is not set, the assistant will be created automatically. You can use the assistant ID to connect to an existing assistant, which can be found in the OpenAI platform after creating an assistant.

Create a .env file in the root directory of your project and populate it with the necessary secrets:

touch .env

Add the following content to the .env file:

# OpenAI API Key
OPENAI_API_KEY=

# Assistant ID - leave it empty if you don't have an assistant to reuse
ASSISTANT_ID=

Please note that the .env file should not be committed to the repository. Add the .env file to the .gitignore file to prevent it from being committed.

This was the first step needed to run the library. The next step is to configure the assistant.

The library provides a way to configure the assistant with the AssistantModule.forRoot method. The method takes a configuration object as an argument. Create a new configuration file like in a sample configuration file (chat.config.ts) and fill it with the necessary configuration.

// chat.config.ts file
import { AssistantConfigParams } from '@boldare/openai-assistant';
import { AssistantCreateParams } from 'openai/resources/beta';

// Default OpenAI configuration
export const assistantParams: AssistantCreateParams = {
  name: 'Your assistant name',
  instructions: `You are a chatbot assistant. Speak briefly and clearly.`,
  tools: [{ type: 'file_search'}],
  model: 'gpt-4-turbo',
  temperature: 0.05,
};

// Additional configuration for our assistant
export const assistantConfig: AssistantConfigParams = {
  id: process.env['ASSISTANT_ID'],
  params: assistantParams,
  filesDir: './apps/api/src/app/knowledge',
  toolResources: {
    file_search: {
      // Provide files if you use file_search tool
      fileNames: ['example1.txt', 'example2.txt'],
    },
  },
};

More details about the configuration can be found in the wiki.

What is this step for?

From now you can run your application and call the assistant.

Function calling allows you to extend the assistant’s capabilities with custom logic. If you are not going to use function calling you can jump to: Step 5: Testing.

Create a new service that extends the AgentBase class, fill the definition and implement the output method.

• The output method is the main method that will be called when the function is invoked by the assistant.
• The definition property is an object that describes the function and its parameters so the assistant can understand how to call it.

For more information about function calling, you can refer to the OpenAI documentation.

The instructions for creating a function can be found in the wiki, while examples can be found in the agents directory.

What is this step for?

If you’ve defined a function and the output method, you can now call it from the assistant just by asking him to do the action described in the function definition.

Run your application and this will allow you to test the assistant.

# use this if you are using the repository:
npm run start:dev

# if you are using your own NestJS application, please check the npm scripts in the package.json file
# defualt command for NestJS is:
npm run start

Then you can test the assistant.

1. First, you need to create a thread. You can create one by sending a POST request to the /assistant/threads endpoint with the empty object in the body.
2. Then you can send a message to the assistant by sending a POST request to the /assistant/chat endpoint with the following body:

{
  "threadId": "your-thread-id",
  "content": "Hello, how are you?"
}

3. The assistant will respond with a message. You can send more messages to the assistant by sending a POST request to the /assistant/chat endpoint with the same body as in step 2.

Congrats! You have successfully integrated the AI Assistant library into your NestJS application. 🎉