Skip to content

Latest commit

 

History

History
53 lines (41 loc) · 2.97 KB

README.md

File metadata and controls

53 lines (41 loc) · 2.97 KB

About

The AI-Powered Slack Bot integrates Hugging Face's models into Slack, enabling AI-driven conversations within your workspace. Built with Python, it uses the Slack SDK and OpenAI's API to facilitate natural language interactions.

Stack

python flask huggingface slack redis ngrok docker git guthub

How to run

Create an app

  • at first you need to create and app with the slack app manager
  • with your app created, in "Yours apps" session you will be able to get all theses env vars:

slack_env_vars

slack_app_credentials

  • to be able to install your app in any workspace you need to set an redirect url, in case of testing i just used https://localhost:5000, with this feature we can create an 'add to slack' button to distribute our app, for more details read this

IMPORTANT:

  • Slack only accepts https protocol so you need to create ssl certificates and run in the flask application
  • To handle event subscriptions you need to pass an url in order to enable slacks to making post requests for our app, so ive used ngrok too create a temporary test domain, you only need to worry about running the flask app in https, generate a domain with ngrok, enable events, subscribe bot to the events that you need and voilá...app is not configured yet lol.

AI

  • Create on account in Hugging Face and generate an HF token to put in the env var.
  • its a simple prompt that generate what do you want, u can change the default prompt in llm_service.py

Installing dependencies

python -m venv .venv
source .venv/bin/activate
pip install poetry
poetry install --no-root

Running

  • docker:
docker compose -f .\docker-compose.yml up -d --build  
python backend/app.py