PulseAI

Alerts
Analisis
Auto Scan
Dashboard
Landing
Report
Scans

Track Selected: AI Agents & MCP (Primary) + Automate Your Workflow (Secondary)

The Problem

Every startup needs to know what their competitors are doing. Pricing changes, feature launches, funding rounds, product pivots — missing these signals costs market share. Yet the reality is: manual competitive monitoring takes 5–10 hours per week, enterprise tools like Crayon or Klue cost $500–$2,000/month, and most early-stage teams simply can't afford either the time or the money. The result is that most startups make strategic decisions based on outdated or incomplete competitive data.

The Solution

PulseAI is an autonomous competitive intelligence engine. You give it your product category and up to 5 competitor URLs. PulseAI handles everything else: scraping, researching, analyzing, and delivering structured intelligence to your team — automatically, every single week.

How It Works

PulseAI's multi-agent pipeline runs in four stages:

Stage 1 — Autonomous Data Collection: A Scrape Agent uses Firecrawl MCP to extract pricing plans, features, and blog posts from each competitor's website. Simultaneously, a Research Agent uses Exa MCP to search the web for news, press releases, product launches, funding rounds, and community reviews.
Stage 2 — AI Synthesis: A DeepSeek-powered Analysis Agent (via Featherless AI) synthesizes all raw data into structured competitive intelligence: SWOT analysis, pricing breakdown, feature gap analysis, executive summary, and a 1–10 threat score. A Strategy Agent then generates 5 specific, actionable recommendations for how to compete.
Stage 3 — Multi-channel Delivery: Reports are saved to a real-time Convex database. A formatted digest is posted to Slack, a full report is emailed via Resend, a Notion page is created, and a Google Sheets row is appended for audit trail — all simultaneously.
Stage 4 — Smart Alerting: An If-Else node monitors every report for high-threat signals. If a competitor dropped prices, launched a major feature, or received a threat score ≥ 8, PulseAI fires a critical ⚠️ alert to Slack immediately.

The Cron node runs this entire pipeline every Monday at 8 AM without any manual intervention.

The dashboard provides a real-time view of all reports as cards, a threat activity chart across competitors, and "Ask PulseAI" — a chat interface where you can ask questions like "Which competitor is the biggest threat this month?" and get instant answers from your stored intelligence data.

What Makes It Unique

PulseAI is not a demo — it's a functional autonomous SaaS built in 48 hours. The multi-agent architecture (4 specialized AI agents with different roles), real-time Convex persistence enabling trend tracking over time, multi-channel output delivery, and the "Ask PulseAI" conversational interface combine to create something that goes far beyond typical hackathon submissions.

Built With VibeFlow

PulseAI uses virtually every VibeFlow capability: Agent nodes (×4), MCP nodes (Exa + Firecrawl), Webhook, Collection, Query, Mutation, Cron, For Loop, If-Else, Code, Edit Fields, Slack, Email (Resend), Notion, Google Sheets, Frontend Element, and Return nodes — deployed as a full-stack application with a real-time frontend.