Janus: Your Agentic GTM OS — at Scale

Janus: an AI-native Go-To-Market OS for technical founders. A canvas with agents that plan, A/B test, and reroute soc-med posts in real time from metrics—cutting CAC and ship your marketing at scale.

Comment

Inspiration

As technical founders, we’ve all faced the same pain: we love building products but hate doing marketing. Hiring marketing managers is expensive, and managing social media manually is chaotic. We realized the true bottleneck wasn’t content creation—it was iteration speed. So we built Janus, an AI-native Go-To-Market OS that plans, learns, and adapts like a marketing team on autopilot.

What it does

Janus automates your entire go-to-market strategy. It visualizes campaigns as nodes on a canvas, dynamically changes posts based on real-time engagement metrics, and continuously re-optimizes your marketing direction. It connects APIs from X (Twitter), Instagram, and Product Hunt to gather data, then re-writes and re-routes content — all with your approval. In short: An AI-native GTM OS that learns from your metrics and adapts your strategy in real time.

How we built it

  • Frontend: React + Tailwind for the Canvas OS UI
  • Graph Engine: ReactFlow + Mermaid parsing to visualize marketing phases, posts, and campaigns
  • Backend: Django REST Framework + LangChain + Gemini API for multi-agent orchestration
  • Data Layer: X, Instagram, and Product Hunt APIs for metric collection
  • We modeled Janus as a dynamic graph of marketing nodes and feedback edges updated by AI.

Challenges we ran into

  1. Real-time node updates without breaking the graph structure
  2. API rate limits while pulling engagement metrics
  3. Designing a feedback loop that felt intuitive but gave founders full control
  4. Time — building a working AI-driven UX within 48 hours

Accomplishments that we're proud of

  • Built a working AI-native Canvas OS that visualizes and updates marketing flows in real time
  • Integrated A/B testing nodes with automatic metric-based adaptation
  • Created a human-in-the-loop approval system for safe AI automation
  • Designed an intuitive metrics dashboard for startup founders

What we learned

Automation doesn’t replace human insight — it amplifies it. By combining AI iteration with human supervision, founders can maintain creative control while scaling execution. We also learned how to manage complex node-based UIs, integrate multiple APIs efficiently, and design adaptive feedback systems.

What's next for Janus

We plan to expand beyond hackathon scope:

  • Launch as startup
  • Integrate LinkedIn and TikTok APIs for broader marketing coverage
  • Build more adaptive triggers (e.g. “if ER < 2%, generate new creative”)
  • Open-source Janus to support early-stage founders globally
  • Develop a premium “AI Marketing OS” SaaS version for startups

Revenue Model for Janus

Our main business and revenue model for Janus will be a B2B SaaS model. We plan on having multiple product lines, supporting technical founders at each stage of their startup journey:

  • Free Plan: Indie Builders, 500 credits/month, X.com, $0/month
  • Starter Plan: Pre Seed, 10000 credits/month, X.com/Instagram/ProductHunt, $50/month
  • Pro Plan: Seed A, 50000 credits/month, X.com/Instagram/ProductHunt/TikTok, $200/month

Built With

Share this project:

Updates