Narrative AI

The Challenge

Professionals and businesses struggle to maintain consistent, high-quality content on LinkedIn. Creating engaging posts requires significant time for research, writing, and optimization—time that most don't have. Generic AI tools produce shallow, formulaic content that fails to establish thought leadership.

The Innovation

Narrative AI reimagines content creation through a multi-agent system that mimics how expert content creators work. Built on the Agno Framework v2.0, the platform employs specialized AI agents for research, writing, editing, and optimization—each powered by Google's Gemini AI.

Agent Architecture

Research Agent

• Integrates Brave Search API for web research
• Jina Async Scraper for data extraction
• Grounds posts in current events, industry trends, and data-backed insights

Writing Agent

• Crafts narratives that balance professional insight with conversational tone
• Structured output generation using Pydantic models
• Maintains authentic voice while optimizing for engagement

Editor Agent

• Refines for clarity, engagement, and platform-specific best practices
• Professional post formatting with smart use of emojis and hashtags
• Ensures content feels authentically "human"

Beautiful Agent UI

Built with Next.js + TypeScript, the interface provides:

• 🎨 Real-time visualization of agent collaboration
• 💬 Natural conversation interface
• 💾 Persistent chat sessions with complete history
• 🔄 Agent memory powered by SQLite (with optional Neon/Supabase)
• 📊 Transparent agent reasoning—users see why certain choices were made

Production Features

• Database Options: Local SQLite, NeonDB, or Supabase for session storage
• Environment Flexibility: Easily configurable via .env
• FastAPI Playground: localhost:7777 server for agent interactions
• Developer Experience: Built with modern Python tooling (uv, pyproject.toml)

Real Results

"More details in the prompt = Better output 😉"

Early adopters report 5x faster content creation with significantly higher engagement rates. The transparent agent workflow builds trust—users understand the reasoning behind each suggestion. Most importantly, the content feels authentically theirs, not generic AI text.

Open Innovation

Licensed under PolyForm Noncommercial 1.0.0:

• ✅ FREE for individuals (personal projects, learning, research)
• 💼 Commercial license available for organizations

Note: The project has evolved from a deprecated Streamlit interface to the current Agent UI, offering conversation history, agent memory, and enhanced UX.