SDLC Automation Bot — Meeting to Tasks

Overview

RoyalSoft meeting/SDLC stack: start meeting bots, capture transcripts, run multi-stage AI treatment (summary → requirements → tasks), and sync structured work into TMS — accelerating project kick-off.

Problem

Meeting notes stay unstructured. Turning discussions into SDLC requirements, tasks, and assignments is manual and slow.

Business challenge

Automate kick-off from live meetings without losing audit history or developer assignment accuracy.

Solution

Next.js app drives n8n webhooks for Recall.ai bot lifecycle, Claude treatment pipelines, SDLC execute hooks, and PostgreSQL session history.

Architecture

Next.js UI → n8n webhooks (start/status/leave/transcript/treat/sdlc) → Recall.ai + Claude + TMS APIs → PostgreSQL session_history.

Implementation

TypeScript App Router UI with Zustand state, mammoth DOCX parsing, react-markdown transcript/treatment views, pg history API, and self-hosted n8n on Ubuntu/PM2/Nginx.

Challenges

Coordinating bot lifecycle, realtime transcripts, and reliable task creation across multiple external APIs.

How we solved them

Webhook-oriented n8n design, dedicated treat/sdlc execute endpoints, and persistent session history for replay and audit.

Results

Faster SDLC kick-off from meetings with structured tasks flowing into TMS.

Key modules

• Next.js app
• n8n workflows
• Recall.ai integration
• TMS sync
• History DB

Features

• Dashboard for transcript, treatment, SDLC, history, and automation
• Recall.ai bot start/status/leave/transcript webhooks
• Claude multi-stage treatment pipeline
• TMS developer lookup and task creation
• DOCX upload and markdown rendering
• PostgreSQL-backed session history

My responsibilities

• Built meeting bot control UI and history persistence
• Wired n8n treatment and SDLC execute workflows
• Integrated Recall.ai and TMS APIs

Outcomes

• 60%+ faster project kick-off