Verbatim Vigil
A niche, AI-powered subscription service that scrapes and analyzes public municipal documents (e.g., city council minutes, ordinances) based on user-defined rules, delivering high-value, early-warning alerts to professionals.
Memoria Engine
A smart document management tool for researchers and journalists that automatically assembles fragmented information from various sources into coherent, explorable timelines and narratives.
Inspired by the fragmented memory in 'Memento', the assemblage of parts in 'Frankenstein', and the automated data gathering of a 'Weather Forecasts' scraper, Memoria Engine is a niche document management system designed to reconstruct narratives from chaos.
The Story: Like 'Memento's' protagonist, investigative professionals (journalists, academics, legal analysts) rely on a trail of documents—interview transcripts, articles, emails, PDFs—to piece together the truth. Memoria Engine acts as their structured memory system. Instead of simply storing files in folders, it helps them 'Frankenstein' a coherent body of evidence from disparate limbs of information, creating a living, explorable story from dead data.
The Concept: The project shifts the paradigm from static, location-based file storage to dynamic, context-based narrative construction. The core of the system is not a file tree, but an interactive timeline and relationship graph. It's designed for users who need to answer complex questions like, 'What sequence of events led to this outcome?' or 'How are these three individuals connected across all my source material?'
How it Works:
1. Ingestion & Scraping: Users connect their data sources (e.g., Google Drive, Dropbox folders, Zotero libraries, specific web URLs). The system acts like a persistent scraper, automatically ingesting documents and web content, extracting raw text and metadata.
2. Fragmentation & Annotation: Using NLP, the engine automatically identifies and tags key entities (people, dates, locations, organizations) within every document. Users can also manually highlight and annotate crucial snippets of text or data, creating 'memory fragments'—the core units of the system, much like the photos and tattoos in 'Memento'.
3. Assembly & Visualization: The 'memory fragments' are plotted on an interactive timeline and a visual graph. The AI suggests potential connections between fragments based on shared entities, semantic similarity, and chronological proximity. Users can then manually 'stitch' these fragments together, building arguments, mapping networks of influence, and reconstructing event timelines.
4. Forecasting Gaps: Drawing inspiration from weather forecasting's predictive nature, the engine analyzes the constructed narrative to identify logical or chronological 'gaps' in the evidence. It can then suggest areas for further research, helping the user know what they don't yet know and where to look next.
The Story: Like 'Memento's' protagonist, investigative professionals (journalists, academics, legal analysts) rely on a trail of documents—interview transcripts, articles, emails, PDFs—to piece together the truth. Memoria Engine acts as their structured memory system. Instead of simply storing files in folders, it helps them 'Frankenstein' a coherent body of evidence from disparate limbs of information, creating a living, explorable story from dead data.
The Concept: The project shifts the paradigm from static, location-based file storage to dynamic, context-based narrative construction. The core of the system is not a file tree, but an interactive timeline and relationship graph. It's designed for users who need to answer complex questions like, 'What sequence of events led to this outcome?' or 'How are these three individuals connected across all my source material?'
How it Works:
1. Ingestion & Scraping: Users connect their data sources (e.g., Google Drive, Dropbox folders, Zotero libraries, specific web URLs). The system acts like a persistent scraper, automatically ingesting documents and web content, extracting raw text and metadata.
2. Fragmentation & Annotation: Using NLP, the engine automatically identifies and tags key entities (people, dates, locations, organizations) within every document. Users can also manually highlight and annotate crucial snippets of text or data, creating 'memory fragments'—the core units of the system, much like the photos and tattoos in 'Memento'.
3. Assembly & Visualization: The 'memory fragments' are plotted on an interactive timeline and a visual graph. The AI suggests potential connections between fragments based on shared entities, semantic similarity, and chronological proximity. Users can then manually 'stitch' these fragments together, building arguments, mapping networks of influence, and reconstructing event timelines.
4. Forecasting Gaps: Drawing inspiration from weather forecasting's predictive nature, the engine analyzes the constructed narrative to identify logical or chronological 'gaps' in the evidence. It can then suggest areas for further research, helping the user know what they don't yet know and where to look next.