Inspiration

While conducting their core business, local companies are expected to act as full-fledged marketing departments, posting material every day, monitoring rivals, reacting to clients immediately, and analyzing performance data. The majority of tools just automate tasks.
Automating decision-making was our goal.

Viralis was founded on a straightforward concept: What if an AI could manage the entire growth engine of a company?

Viralis is an independent social growth CRM that is powered by Google Gemini and is made to plan, think, and act on behalf of a company.

What it does

Viralis is a growth platform powered by AI that enables companies to expand their online presence with little manual labor.

It allows companies to:

  • Create graphic concepts, hooks, video scripts, and platform-aware content strategies.
  • Evaluate rivals and get practical, instantaneous growth suggestions.
  • Convert performance data into understandable, actionable insights.
  • Use a low-latency artificial intelligence voice assistant to answer consumer calls.
  • Keep your brand's voice consistent throughout all of your content.
  • Improve messaging for local search engine optimization and discovery.

Viralis doesn’t just create content—it coordinates growth.

How we built it

Viralis is an agent-based, modular system built around Google Gemini.

  • Content strategy, script creation, competitive analysis, and growth reasoning are all powered by Gemini Pro.
  • Real-time speech engagements with sub-second reaction times are made possible using Gemini Flash.
  • Gemini Embeddings provide long-term brand memory by learning from historical content and conversations

TypeScript and Node.js are used in the platform's construction, and workflows for voice, analytics interpretation, competitive intelligence, and content creation are orchestrated via APIs.
For non-technical users, the interface provides a targeted, decision-first dashboard.

Challenges we ran into

The team encountered multiple challenges, which included:

  • Designing prompts that work with many platforms and types of information
  • Maintaining a consistent brand voice across several AI agents
  • Low-latency answers are necessary for the system to function and allow users to communicate by voice in real time.
  • The procedure for turning noisy analytics into suggestions that may be put into practice.
  • Working within the constraints of the hackathon, the team had to develop a full-production system.

Each challenge reinforced the importance of system design over isolated AI features.

Accomplishments that we're proud of

  • Coordinated several Gemini models for specific tasks
  • An AI was created that provides decisions rather than dashboards.
  • Context-aware, real-time voice agent implementation
  • Under schedule constraints, a scalable and adaptable architecture was shipped.

What we learned

Orchestration, not generation, is the most challenging aspect of AI products. When AI explains why rather than just what, users are more likely to trust it. Since latency serves as a component of user experience design, conversational systems' user experience is dependent on it.

What's next for VIRALIS

  • Automated video rendering and multi-platform publishing
  • Multi-language and multi-region expansion
  • Deeper CRM and booking integrations
  • The team needs to transform Viralis into a SaaS platform that meets all production requirements.

Viralis is built to scale beyond a hackathon—into a real-world growth engine.

Built With

Share this project:

Updates