Marketing-Agent

Landing View with proper CTA's
Brand Settings Page
Tasks that are being created after analysing website
Content Creation Dashboard
Brand Analytics Page
Main Analytics Dashboard
Responsive Navigation
Proper Navigation items
Blogs Page - Main section
More Blogs and Filtering Option
Product Documentation
Visualisation of Product in documentation
Help Center

Inspiration

Marketing is the lifeblood of any business, but for small teams and solo founders, maintaining a consistent brand voice and churning out high-quality content daily is exhausting. We noticed that while AI tools exist, they often feel "robotic" or require endless prompting to get the tone right. We were inspired to build Marketing Agent—not just another text generator, but a strategic partner that truly understands a brand's DNA, goals, and audience, automating the workflow from strategy to execution.

What it does

Marketing Agent is an intelligent dashboard that acts as a dedicated CMO (Chief Marketing Officer) in your pocket.

Brand Identity Engine: Users input their core values, audience personas, and voice settings (witty, professional, etc.), which the agent "memorizes" to ensure consistency.
Campaign Wizard: A guided flow where users state a simple goal (e.g., "Drive sales for our summer launch"), and the agent generates a comprehensive multi-channel campaign.
Smart Content Generation: Automatically creates blog posts, social media captions, and ad copy that align perfectly with the user's specific brand guidelines.
Competitor Awareness: Contains safety checks to avoid forbidden words or mentioning competitors.

How we built it

We built the platform using Next.js 14 (App Router) for a responsive and modern frontend, styled with Tailwind CSS.

Backend: We used Supabase for authentication and as our primary database to store complex brand profiles and campaign data.
AI Integration: The core intelligence is powered by LLMs (via Supabase Edge Functions), which ingest the structured brand data we collect to generate highly contextualized outputs.
Database Design: We implemented a robust schema to handle relational data between users, brands, and their specific "voice" settings.

Challenges we ran into

The biggest challenge was "context stuffing." Passing enough data to the AI to make it sound exactly like the user's brand without hitting token limits was tricky. We solved this by creating a refined "Brand Persona" object that condenses key attributes (tone, forbidden words, emojis) into a dense system prompt. Another hurdle was the Wizard UI state management—ensuring the multi-step form felt seamless and retained data if the user navigated away. We utilized local state and efficient database syncing to solve this.

Accomplishments that we're proud of

We are particularly proud of the Brand Settings Editor. It’s not just a form; it’s a comprehensive identity builder that captures nuances like "Emoji Usage" and "Tone of Voice," which dramatically improves the quality of the AI's output. Seeing the AI generate a campaign that actually sounded like a human wrote it was a huge win.

What's next for Marketing Agent

Auto-Posting: connecting directly to Instagram/LinkedIn APIs to schedule the generated content.
Visual Generation: integrating image generation models to create on-brand visuals alongside the copy.
Analytics Loop: feeding engagement data back into the system so the agent learns what content performs best over time.