Invenio

Product Introduction

1. Definition: Invenio is a local-first, AI-powered video and photo search application for macOS. It falls into the technical categories of Media Asset Management (MAM), intelligent video retrieval, and on-device machine learning software.
2. Core Value Proposition: Invenio exists to eliminate the time-consuming, manual process of searching through large video and photo libraries. Its primary value is enabling instant, natural language search across visual content, spoken dialogue, and on-screen text, with a foundational commitment to 100% local AI processing and offline-first privacy. This addresses the critical need for fast video editing workflows and secure media management for professionals.

Main Features

1. Semantic Visual Search: This feature allows users to find video clips and photos by describing the scene in natural language. It works by using on-device computer vision models (optimized for the Apple Neural Engine) to analyze and index every frame of your media. The AI understands objects, actions, scenes, compositions, and colors, creating a semantic map of your library. When you search for "man on a bike" or "drone shot of mountains at sunset," it matches your query against this visual understanding, not just filenames or metadata.
2. Speech & Dialogue Transcript Search: This feature automatically transcribes all spoken words within video files, including dialogue, voiceovers, and interviews. It utilizes local AI speech recognition models to create a searchable index of every uttered word. This allows for precise retrieval of clips based on spoken content, such as finding the exact moment a client said "Q3 Revenue Growth" or an interviewee mentioned a specific phrase, searching spoken words across thousands of videos.
3. On-Screen Text Recognition (OCR): This feature employs Optical Character Recognition (OCR) AI to detect and read text that appears visually within video frames and photos. It indexes text from slides, subtitles, street signs, documents, and whiteboards. Users can then search for this text, enabling use cases like finding a lecture slide titled "Introduction to Biology" or a scene with a specific street name, making OCR search a powerful tool for educational and documentary footage.
4. Local-Only, Offline-First Architecture: A foundational feature, not an afterthought. All AI processing—visual analysis, speech transcription, and OCR—is executed locally on the user's Mac. It leverages Apple Silicon's Neural Engine for efficient, high-performance computation. No media files, metadata, or transcripts are ever uploaded to the cloud. This ensures professional-grade privacy, offline video search capability, and compliance with NDA and sensitive project requirements.

Problems Solved

1. Pain Point: The immense time cost and creative disruption of manually scrubbing through hours or terabytes of raw video footage to find a specific shot, known as "video logging." This process wastes billable hours and breaks creative flow during editing.
2. Target Audience: The primary user personas are professional video editors, documentary filmmakers, YouTubers, content creators, and freelance colorists. Secondary audiences include academic researchers, journalists, and podcast editors who need to search through large archives of recorded lectures, interviews, or audio-visual assets.
3. Use Cases:
   • A documentary filmmaker needing to locate a specific interview clip from a 4TB archive shot over years, based on a remembered topic of discussion.
   • A YouTuber quickly finding B-roll of a "red car in desert" from footage shot three years prior to complement a new project.
   • A commercial editor working under an NDA who must search client footage without any risk of data leaving their secure machine.
   • A researcher transcribing and searching for specific terminology across hundreds of hours of recorded lectures and presentation slides.

Unique Advantages

1. Differentiation: Unlike cloud-based AI media tools (like Google Photos, Adobe Sensei features requiring uploads), Invenio processes everything locally, offering superior privacy and speed for large files. Compared to traditional media managers (like Eagle, Adobe Bridge) or system search (Spotlight), it searches by understanding content rather than relying on manually applied tags, filenames, or basic metadata.
2. Key Innovation: The seamless integration of multiple, optimized on-device AI models (visual, speech, OCR) into a single, fast retrieval system that works entirely offline. The specific engineering to make this feasible and performant on consumer Apple Silicon hardware, handling terabytes of footage with instant retrieval, is its core technical innovation. The drag-and-drop integration directly into NLE timelines (Premiere Pro, Final Cut Pro, DaVinci Resolve) from the menu bar is a key workflow innovation.

Frequently Asked Questions (FAQ)

1. How does Invenio's AI search work without the internet? Invenio uses AI models that are downloaded and stored directly on your Mac. All processing—analyzing video frames, transcribing speech, reading text—is performed locally by your computer's hardware, specifically the Apple Neural Engine, requiring no cloud connectivity or data uploads.
2. Is Invenio compatible with external hard drives and large video libraries? Yes, Invenio is built for professional-scale media management. It fully supports indexing and searching media stored on external SSDs, HDDs, and RAID arrays. It is designed to handle libraries spanning terabytes and hundreds of thousands of files with fast, millisecond-level search results.
3. What are the system requirements for running Invenio on Mac? Invenio requires a Mac with Apple Silicon (M1, M2, M3, M4, or later chips) and macOS. It is optimized for the Neural Engine in these chips and does not support Intel-based Macs due to the high-performance demands of the local AI processing.
4. What's the difference between the Free and Pro versions of Invenio? The Free version includes unlimited visual semantic search, on-screen text (OCR) search, support for external drives, and drag-and-drop functionality. The Pro version adds the crucial Speech Transcript Search feature, enabling you to search by spoken dialogue and words within your videos, along with priority support.
5. How long does it take to index a large video collection? Indexing speed is highly optimized for Apple Silicon. While the initial index of a very large library will take time, the process is designed to be efficient and run in the background without significantly impacting system performance or overheating your Mac, allowing you to continue other work.