Meeting Insights Agent

AI-powered Slack bot that turns meeting chaos into instant summaries, action items, and answers in seconds.

Comment

Inspiration

Teams waste hours every week scrolling through long Slack channels to find meeting decisions, action items, and key discussions. With remote and hybrid work, critical meeting insights get buried in endless message threads. We wanted to build something that turns this chaos into clarity — instantly.

What it does

Meeting Insights Agent is an AI-powered Slack bot that analyzes meeting conversations and delivers instant intelligence through three core features:

/meeting-summary — Generates a concise AI summary of recent channel messages, highlighting key topics, decisions, and context. /action-items — Extracts and organizes tasks with assignees, deadlines, and priority levels for clear accountability. @Mention Q&A — Ask natural language questions like "What was decided about dark mode?" and get context-aware, sourced answers from conversation history. The bot automatically detects meeting-related conversations and provides intelligent responses in under 2 seconds.

How we built it

We built the bot using the Slack Bolt Framework with Node.js, connected via Socket Mode (WebSocket) for real-time bidirectional communication — no public endpoint needed. For AI intelligence, we integrated Groq's LLaMA 3.3 70B model for sub-second inference on meeting analysis and Q&A. The entire application is hosted on Render's free tier with UptimeRobot pinging every 5 minutes to keep the instance awake, ensuring 99.9% availability. The smart context window fetches up to 200 recent messages from Slack's Web API for accurate AI responses. Total cost: $0/month.

Challenges we ran into

The biggest challenge was managing AI context within Slack's API rate limits and message fetch constraints. We had to implement a smart pagination system to gather enough conversation history for accurate summaries without hitting rate limits. Another challenge was ensuring the bot's responses were contextually relevant — we solved this by crafting precise system prompts and tuning the context window size for optimal results.

Accomplishments that we're proud of

Built a fully functional AI Slack bot with zero recurring cost ($0/month) Achieved sub-2-second response times using Groq's ultra-fast LLM inference Implemented three distinct interaction patterns: slash commands, @mention Q&A, and auto-detection Deployed a production-ready bot with 99.9% uptime using free-tier tools only Created an intelligent context window that delivers accurate, sourced responses

What we learned

We learned how to effectively integrate LLM capabilities into real-time messaging platforms. Slack's Bolt Framework and Socket Mode made it surprisingly easy to build bidirectional communication. We also learned the importance of prompt engineering — crafting the right system prompts dramatically improved response quality. Additionally, we gained hands-on experience with Groq's API and the power of fast LLM inference for productivity tools.

What's next for Meeting Insights Agent

Thread-level analysis — Summarize and extract action items from specific threads Scheduled digests — Automatic daily/weekly meeting summary posts Multi-channel support — Aggregate insights across multiple Slack channels Integration with project management tools — Sync action items to Jira, Notion, or Linear Sentiment analysis — Detect meeting tone and team morale trends over time

Built With

Share this project:

Updates