Gencite

• 

  Dark mode ui 2

• 

  content ideas generated for users to create content around these topics to bridge gaps in search queries

• 

  Platform generate content ideas to bridge the gaps in search query related content

• 

  Enter the AI search query you want to analyze

• 

  Input your brand / company name

• 

  Dark mode ui

• 

  Generate human like variations of this queries

• 

  Running query analysis through Chatgpt & Gemini

• 

  Analyze share of voice results

• 

  See citations by search query

## Inspiration

Brands increasingly depend on AI assistants to shape how customers discover products, compare options, and make decisions. Traditional SEO tools measure search rankings, but they rarely show whether ChatGPT or Gemini actually mentions a brand. We built GenCite to make that visibility measurable and actionable.

## What it does

GenCite measures a brand’s share of voice across AI-generated answers. Users enter a topic and brand, select realistic customer-query variations, and run them through ChatGPT and Gemini. GenCite then shows mention rates, competitor gaps, model-by-model performance, and new keyword recommendations to test next.

## How we built it

We built the interface with React, TypeScript, and Vite. Firebase provides Google authentication, Firestore, and callable Cloud Functions. The backend sends structured requests to OpenAI and Gemini, runs both providers in parallel, validates their responses, and aggregates the results into a unified dashboard. A recommendation engine uses measured query gaps to suggest new AI-search opportunities.

## Challenges we ran into

The biggest challenge was turning nondeterministic AI answers into reliable, comparable metrics. We needed repeated samples, strict structured-output validation, consistent query handling, and graceful fallbacks when a provider failed or reached its quota. We also had to distinguish genuine brand mentions from noisy model output while keeping API keys and analysis logic secure on the server.

## Accomplishments that we're proud of

We are proud that GenCite goes beyond generating content: it creates a repeatable measurement workflow. It compares multiple AI providers side by side, preserves results across repeated runs, surfaces competitor gaps, and connects every recommendation back to measured evidence. The system also continues working when one provider is unavailable.

## What we learned

We learned that visibility in generative search is highly dependent on phrasing, intent, and model choice. A brand may perform well for direct questions but disappear from comparisons, pricing questions, or audience- specific searches. We also learned that trustworthy AI analytics requires deterministic validation and transparent evidence—not just another model- generated score.

## What's next for GenCite

Next, we plan to improve recommendation ranking with stronger evidence- based scoring, semantic deduplication, and feedback from previously tested suggestions. We also want to add historical trend dashboards, scheduled monitoring, richer citation-source analysis, team workspaces, exports, and integrations with search-demand data. Our goal is to make GenCite the control center for understanding and improving how brands appear across AI search.

Built With

• chatgpt
• codex
• css
• firebase
• gemini
• html5
• playwright
• react
• typescript
• vite