A Simple Slack Bot developed using Python and slack_bolt that interacts with users in a Slack channel, providing responses powered by an external API that utilizes LlamaIndex and AI to process natural language from documents.
The bot allows you to chat with an AI assistant powered by any AI provider with documentation injected with LlamaIndex directly within a Slack workspace. It shows how retrival augmented generation can be integrated into an existing chat platform.
- Responds to events in a defined Slack channel.
- Engages with users through direct messages and @mentions.
- Maintains communication in a thread and avoids responding to its own messages.
In the above example I have added my CV into the documents folder, then asked questions about it.
Clone this repo Ensure OPENAI_API_KEY, SLACK_BOT_TOKEN and SLACK_SIGNING_SECRET are in .env file
pip install -r requirements.txt
python server.pyPull requests are welcome to extend the bot's functionality.
- In case of a 404 error from the API, the bot will notify the user with an error message.
- In instances of API failure, the bot will wait 1 second and attempt the request again.
- Deployment strategy and hosting platform are under consideration (suggestions welcomed).
- Monitor git pushes and restarts the bot as necessary, primarily during off-hours unless urgent.
- Implement Prompt Injection prevention to safeguard from potential hijacking.
- Implement a rate limiter to prevent abuse.