How to Integrate OpenAI with Slack

Prerequisites and Setup

• Create an OpenAI account and obtain an API key. This key will be crucial for accessing OpenAI services.

 

• Create a Slack account and set up a workspace if you haven't already. You will also need permissions to add and configure applications in your Slack workspace.

 

• Ensure you have basic programming knowledge and a local development environment set up to write and test your code.

Create a Slack Bot

• Navigate to the Slack API page at https://api.slack.com/.

 

• Click on “Create an App” and choose “From scratch.” This will create a new bot for your workspace.

 

• Enter a name for your bot and select the workspace where you want to install it, then click “Create App.”

 

• Under “Add features and functionality,” choose “Bots.” Click on “Review Scopes to Add” under Bot Token Scopes and include necessary scopes like `chat:write` and `im:history`.

Install the Bot in Your Workspace

• In the left sidebar, find and click on “Install App.”

 

• Click on “Install to Workspace,” and authorize the bot by following the prompts. You'll receive a Bot User OAuth Access Token; keep this token handy as you'll need it for integration later.

Set up a Project Environment

• On your local machine, create a directory for your project. Navigate into it and initialize a new Node.js project:

mkdir my-slack-openai-bot
cd my-slack-openai-bot
npm init -y

• Install the necessary packages for making HTTP requests and connecting to Slack:

npm install @slack/web-api axios

Write Code to Integrate OpenAI API with Slack

• Create a file called `bot.js` in your project directory.

 

• Add the following code to connect to both OpenAI and Slack:

const { WebClient } = require('@slack/web-api');
const axios = require('axios');

// Initialize Slack client with your bot token
const slackToken = 'YOUR_SLACK_BOT_TOKEN';
const slackClient = new WebClient(slackToken);

// Your OpenAI API key
const openaiApiKey = 'YOUR_OPENAI_API_KEY';

// Listen for messages in a specific channel
async function listenToChannel(channelId) {
  const result = await slackClient.conversations.history({ channel: channelId });
  for (const message of result.messages) {
    // Process and send each message to OpenAI
    if (message.text) {
      const response = await callOpenAI(message.text);
      await slackClient.chat.postMessage({
        channel: channelId,
        text: response.data.choices[0].text,
      });
    }
  }
}

// Function to call OpenAI API
async function callOpenAI(prompt) {
  return await axios.post('https://api.openai.com/v1/engines/davinci-codex/completions', 
  {
    headers: { 
      'Authorization': `Bearer ${openaiApiKey}` 
    },
    data: {
      prompt: prompt,
      max_tokens: 150
    }
  });
}

// Replace 'CHANNEL_ID' with your Slack channel id
listenToChannel('CHANNEL_ID');

• Replace `YOUR_SLACK_BOT_TOKEN`, `YOUR_OPENAI_API_KEY`, and `CHANNEL_ID` with your Slack bot token, OpenAI API key, and the ID of the Slack channel you want the bot to listen to, respectively.

Run Your Slack Bot

• Execute your `bot.js` script using Node.js:

node bot.js

• Once running, the bot will listen to messages in the specified Slack channel and route them to the OpenAI API. The responses can be configured and modified as needed to fit your use case.