-
-
AI-powered dashboard showing meetings, extracted tasks, completion status, and active risks in real time.
-
Upload meeting audio or transcripts and let Gemini 3 convert conversations into structured work items.
-
Automatically generated tasks organized into a Kanban board for tracking progress from discussion to execution.
-
Gemini 3 analyzes meeting transcripts to generate structured summaries and extract actionable tasks automatically.
Inspiration In today’s fast-paced work environment, meetings generate valuable discussions—but execution often breaks down afterward. Teams spend hours aligning on decisions, only to lose track of action items once the call ends. Manual note-taking is time-consuming, error-prone, and rarely actionable. We wanted to solve this gap by building a system that doesn’t just capture meetings, but actively turns conversations into execution-ready work. MeetFlow AI was created to eliminate post-meeting friction and help teams move seamlessly from discussion to delivery.
What it Does MeetFlow AI is an AI-powered productivity platform that converts meeting conversations into a structured project management workspace. Audio & Text Input: Users can upload meeting audio files (MP3, WAV) or paste transcripts directly. AI Task Extraction: Powered by Gemini 3, the system analyzes conversations to identify actionable tasks, infer ownership, determine priorities, and generate realistic deadlines. Automated Kanban Board: Extracted tasks are instantly organized into a Kanban workflow (Todo, In Progress, Done) for immediate visibility and execution. Smart Follow-ups: MeetFlow AI generates professional follow-up messages and WhatsApp reminders with one click. Risk Analysis: The AI proactively flags potential risks and delays based on task context and dependencies.
How We Built It MeetFlow AI was built using a modern, scalable stack optimized for speed and clarity: Frontend: Vue.js 3 with Vite for a fast, reactive UI, using Pinia for state management. Backend: Laravel 12 provides a secure and extensible API layer. AI Engine: Google Gemini 3 via the Gemini API, leveraging its advanced reasoning, multimodal understanding, and structured JSON output. Database: MySQL for reliable storage of meetings, tasks, and team data. Design: A custom glassmorphism-inspired UI for a polished, modern experience.
Challenges We Ran Into Accurate Task Interpretation: Distinguishing between casual discussion and actionable commitments required careful prompt engineering and iteration. Audio Processing Flow: Managing audio uploads and synchronizing transcription with AI analysis demanded thoughtful backend orchestration. Environment & CORS Configuration: Ensuring secure communication between the Vue frontend and Laravel backend—especially across custom deployment paths—required precise server configuration.
Accomplishments We’re Proud Of End-to-End Automation: Uploading an audio file and instantly receiving a populated Kanban board creates a genuinely magical experience. Production-Grade UI: Despite hackathon constraints, we delivered a clean, intuitive, and professional interface. Deep AI Integration: Beyond basic extraction, MeetFlow AI demonstrates advanced Gemini capabilities such as risk detection and structured follow-up generation.
What We Learned Multimodal AI is Transformational: Gemini’s ability to process both audio and text simplified our architecture and improved accuracy. Context Is Everything: High-quality, structured prompting—especially JSON-based outputs—was critical for reliably integrating AI results into application logic.
What’s Next for MeetFlow AI Live Meeting Integrations: Real-time transcription from Google Meet or Zoom. Voice Commands: Hands-free task creation and updates. Mobile App: Native mobile experience for on-the-go execution. Modular Workflow Extensions: Optional plug-in modules that allow teams to export or sync tasks with tools like Slack, Jira, and other workflow platforms.
Log in or sign up for Devpost to join the conversation.