Agent DNA Lab

An AI-powered Slack agent that answers questions in real time using Google Gemini, built with Node.js and Slack Bolt SDK.

Comment

Inspiration

I wanted to build something that goes beyond a simple rule-based chatbot. As a beginner frontend developer, I set out to create an AI agent that actually understands and responds naturally inside Slack, turning it into a smarter, more helpful workspace

What it does

Agent DNA Lab is a Slack AI agent that listens to messages in a channel and generates real-time, intelligent responses using Google's Gemini API. Instead of relying on hardcoded replies, it processes each message through an AI model, allowing it to answer questions, explain concepts, and hold natural conversations directly inside Slack.

How we built it

I built the agent using Node.js and the Slack Bolt SDK with Socket Mode, which allows the bot to connect to Slack without needing a public server. The bot listens for messages in a channel, sends them to the Gemini API, and posts the AI-generated response back into Slack in real time. Environment variables were used to securely manage the Slack and Gemini API credentials.

Challenges we ran into

As a beginner, I faced several real challenges: setting up OAuth scopes and permissions correctly, debugging syntax errors in the code, and troubleshooting API errors like invalid keys and rate limits. I also had to switch AI providers midway through development after running into billing limitations, which taught me how to adapt my code to work with a different API.

Accomplishments that we're proud of

Successfully building a fully functional, real-time AI-powered Slack bot from scratch as a beginner frontend developer is something I'm genuinely proud of. Going from only knowing frontend UI design to setting up backend infrastructure, handling API integrations, and debugging real production errors was a huge personal milestone.

What we learned

I learned how to work with the Slack API, manage secrets securely using environment variables, debug real error messages step by step, and integrate an external AI API into a working application. Most importantly, I learned that persistence through errors is a core part of building software.

What's next for Agent DNA Lab

Future plans include adding customizable "personality" controls (friendliness, creativity, and tone) that let users adjust how the AI agent responds, along with support for multiple Slack channels and richer conversation memory.

Built With

Share this project:

Updates