

Technical Details
Lead Mobile Developer
Objective-C, AVFoundation, Core Animation
Java, Android SDK
Text-to-Speech, Audio Processing
Gui.de: The News-to-Video Startup That Was a Decade Ahead
A visionary startup that was too early for its time.
In 2013, I joined Gui.de as the lead mobile developer for a startup with a vision that seemed almost science fiction at the time: automatically transform written news articles into video content using text-to-speech, animated graphics, and intelligent summarization. A decade later, this exact concept has become mainstream with AI-powered video generation, but we were building it when the iPhone 5 was cutting-edge.
The Vision
Gui.de's founders saw a future where busy people could consume news passively—watching auto-generated video summaries while commuting, exercising, or multitasking. The app would:
- Aggregate news from multiple sources
- Automatically summarize articles into key points
- Generate video presentations with text overlays and images
- Narrate content using high-quality text-to-speech
- Allow hands-free consumption of information
Technical Implementation
Building this in 2013 meant working within severe technical constraints:
Text-to-Speech Engine:
- iOS 7's AVSpeechSynthesizer was new and limited
- Custom audio processing for more natural pacing
- Phonetic corrections for proper nouns and technical terms
- Background audio session management for continuous playback
- Core Animation for smooth text animations
- Dynamic layout engine for varying content lengths
- Image extraction and cropping from article sources
- Real-time rendering to video frames
- Memory management for continuous video generation
- Article parsing and cleaning from diverse news sources
- Extractive summarization algorithms
- Keyword extraction for visual emphasis
- Reading time estimation and pacing
iOS App Architecture
The app pushed the boundaries of what iOS could do at the time:
Offline-First Design:
- Aggressive pre-caching of content
- Background video generation
- Seamless online/offline transitions
- Storage management for generated videos
- Custom video player with news-specific controls
- Gesture-based navigation between stories
- Sleep timer and continuous playback modes
- AirPlay support for TV viewing
- Careful memory management on 1GB RAM devices
- Incremental video generation
- Image downsampling and caching
- Battery-conscious background processing
Why It Didn't Succeed
Despite the innovative technology, Gui.de faced challenges that ultimately proved insurmountable:
Too Early for the Market:
- Text-to-speech quality was still noticeably robotic
- Users weren't yet comfortable with AI-generated content
- Video content consumption habits hadn't shifted to mobile
- Bandwidth costs were still a concern for users
- News publishers were wary of content transformation
- Monetization through ads in generated videos was complex
- Competition from traditional news apps was intense
- On-device processing couldn't match the vision
- Cloud processing added latency and cost
- Content quality was inconsistent across sources
Looking Back
Gui.de was a learning experience in being too early to market. The core insight—that people want passive, video-based news consumption—has been validated many times over:
- 2016: Snapchat Discover launches with publisher video content
- 2018: TikTok's algorithm-driven short video takes off
- 2020: AI-generated news summaries become mainstream
- 2023: GPT-4 enables high-quality automated summarization
- 2024: AI video generation reaches production quality
While Gui.de didn't achieve commercial success, the technical challenges I solved—real-time video generation, text-to-speech integration, content processing pipelines—have informed my approach to ambitious mobile projects ever since. And every time I see an AI-generated news video, I think about the small team that was working on the same problem a decade earlier.