Skip to content

yash-writes-code/trendsta

Repository files navigation

1. Overview

Trendsta is an AI-powered content growth platform designed to help creators and digital brands discover trends, generate high-performing content ideas, and make data-driven decisions. It combines trend analysis, automated data pipelines, and AI-generated insights into a unified system that reduces guesswork in content creation.


2. Problem Statement

Content creators face several challenges:

  • Difficulty identifying emerging trends early
  • Lack of consistency in content performance
  • No clear understanding of why content goes viral
  • Heavy reliance on intuition instead of data

Existing tools either:

  • Provide raw analytics without actionable insights, or
  • Offer generic suggestions without context

Additionally:

  • Many users do not have the time to interpret dashboards and analytics
  • Even when insights are available, they are not always easy to act upon

3. Solution

Trendsta addresses these challenges by:

  • Collecting and processing trend signals from multiple sources
  • Using AI to generate:
    • Content ideas
    • Hooks
    • Scripts
  • Providing actionable insights instead of raw data
  • Providing a conversational AI consultant for intuitive interaction
    The system focuses on delivering usable outputs, not just analytics.

4. System Architecture

image ---

5. n8n Workflow (Trend Detection & Scraping)

Trendsta currently uses an n8n workflow as the core automation layer for trend detection and data collection.

Purpose of the Workflow

The n8n workflow is responsible for:

  • Scraping or collecting data from socials
  • Extracting relevant signals (engagement, patterns, formats)
  • Structuring this data for further processing
  • Adds an intelligence layer using LLMs

High-Level Workflow Steps

  1. Trigger

    • Invoked by backend API request or scheduled trigger for a particular user
  2. Data Collection

    • Scrapes data from across socials media platform
  3. Preprocessing

    • Cleans and filters raw data
    • Removes irrelevant entries
  4. Intelligence layer

    • Identifies useful indicators such as:
      • Engagement patterns
      • Repeated formats
      • Viral hooks or structures
    • Generates scripts and hashtags for user
  5. Structuring Output

    • Converts processed data into a consistent format
    • Saves it to relevant tables in the database

Why n8n?

  • Enables rapid prototyping of data pipelines
  • Easy to modify and experiment with workflows
  • Reduces initial backend complexity
  • Allows quick iteration on scraping and trend logic

6. AI Consultant (Conversational Interface)

Trendsta includes a conversational AI consultant that allows users to interact with the system in a natural, chat-based format.

Purpose

  • Provides an alternative to traditional dashboards and analytics
  • Allows users to directly ask questions and receive actionable insights
  • Reduces the effort required to interpret complex data

Capabilities

  • Answers queries based on processed trend and research data
  • Suggests content ideas, hooks, and strategies
  • Explains why certain content performs well
  • Assists in decision-making for content direction and growth

Implementation

  • Uses a Retrieval-Augmented Generation (RAG) approach to generate responses grounded in processed trend data
  • Retrieves relevant insights from the database (generated via n8n workflows)
  • Incorporates context management to maintain conversation continuity
  • Uses memory mechanisms to retain user-specific context across interactions
  • Ensures responses are:
    • Context-aware
    • Data-backed
    • Actionable

This allows the AI consultant to function as a stateful, personalized assistant, rather than a stateless chatbot.

8. Tech Stack

  • Webapp : Next.js
  • Automation Layer: n8n
  • Consultant AI Langchain
  • State Management: Zustand
  • Database: PostgreSQL

9. Future Improvements

  • Replace n8n workflows with a custom LangGraph-based pipeline to improve efficiency
  • Improve scalability and performance of trend detection
  • Multi-platform support (YouTube, LinkedIn, X)

11. Limitations

  • AI output depends on LLM quality
  • Trend detection is not fully real-time
  • Current workflow takes some minutes to complete the analysis

12. Use Cases

  • Content creators (Instagram, YouTube, short-form platforms)

  • Influencers and personal brands

  • AI faceless channels

  • Marketing teams and agencies

  • Growth-focused startups

  • AI-generated content

  • Conversational AI consulting

  • Actionable insights

About

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages