Intellectuals

Your AI Newsroom Companion — Real-Time Conversations, Real-World Context

Comment

Inspiration

We were inspired by the idea of making news more interactive, personal, and accessible. With the rise of AI and virtual avatars, we saw an opportunity to reinvent the way people consume information—transforming traditional news into a dynamic, two-way conversation. Instead of passively watching or reading, what if you could simply talk to the news?

What it does

Our project is a Conversational AI News Anchor—a human-like virtual agent capable of real-time, natural conversation. It presents current world events using reliable sources like RSS feeds and news APIs, while also offering historical context from sources like Wikipedia. Key features:

  • Interactive voice-based conversations
  • Real-time news updates
  • Historical background explanations
  • Seamless topic switching and interruption handling
  • Lifelike avatar integration using TTS/STT systems Think of it as a news anchor you can interrupt, question, and casually talk to—just like having a conversation with a journalist.

How we built it

We combined several technologies to build this system:

  • Speech-to-Text (STT): Converts user voice input into text using tools like Whisper.
  • Natural Language Understanding: Handles intent recognition, topic switching, and follow-up queries with a conversational AI engine.
  • Text-to-Speech (TTS): Converts AI responses into natural-sounding speech using ElevenLabs or similar.
  • News Integration: Pulls real-time data from RSS feeds and news APIs.
  • Historical Context: Uses Wikipedia and knowledge graphs to provide accurate background info.
  • Video Avatar: Integrates a realistic, AI-generated anchor that syncs with the TTS output for a lifelike experience.

Challenges we ran into

  • Ensuring smooth real-time interruption handling and context retention.
  • Balancing information density with conversational clarity.
  • Synchronizing the TTS voice with the avatar lip movements effectively.
  • Managing API rate limits and news data consistency.
  • Building a system that sounds natural and human, not robotic or scripted.

Accomplishments that we're proud of

  • Successfully created a fully interactive news anchor that users can talk to naturally.
  • Integrated live news feeds with contextual educational content.
  • Developed a system that responds dynamically to user interruptions and follow-ups.
  • Built an early prototype of a video avatar that mimics human presentation style.

What we learned

  • Combining real-time conversation with structured news delivery requires a deep focus on user experience and flow.
  • Multimodal AI (speech, language, and video) can dramatically enhance engagement.
  • People expect AI to behave contextually, not just factually.
  • The future of media may lie in conversational journalism, where users can shape the narrative through questions and curiosity.

What's next for Intellectuals

We're excited to continue pushing the boundaries of conversational AI and interactive media. Here's what’s next:

  • Enhanced Avatar Realism: Integrate more expressive, emotionally aware avatars using advanced facial animation and gesture synthesis.
  • Multilingual Support: Expand to support real-time news and conversations in multiple languages for a global audience.
  • Deeper Personalization: Allow users to choose topics of interest, preferred speaking style, and voice personality.
  • Fact-Checking Integration: Add real-time verification of breaking news using trusted sources and AI-powered fact-checkers.
  • Offline/Edge Functionality: Develop lightweight versions for offline access and low-resource environments.
  • Public Beta Launch: Release a public-facing beta where users can interact with the AI journalist directly via web and mobile. Ultimately, we aim to create a trusted, accessible AI companion for news, education, and real-time understanding of the world — something that informs, adapts, and engages like a human would.

Built With

Share this project:

Updates