EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·8d ago·5 min read
Show HN: AI search for every photo and every frame of video on macOS
TL;DR
A new AI search tool for macOS allows users to find any photo or video frame using natural language queries, all processed locally without cloud reliance. It features advanced capabilities like scene segmentation and OCR for text recognition.
✦ Why It Matters
Try this tool today to enhance your media search capabilities on macOS without compromising privacy.
Key Takeaways
- Utilize natural language to search for photos and videos.
- Access all features without creating an account or using the cloud.
- Save frequently used queries as tabs for quick retrieval.
- Leverage OCR and Whisper for advanced text and dialogue searches.
Full Summary
This innovative AI search tool for macOS enables users to locate any photo or video frame by describing memories in plain language. It operates entirely on the user's device, ensuring privacy by avoiding cloud storage or account requirements.
The tool employs a local vision model to interpret queries and segments video into scenes, allowing users to jump directly to specific moments. Additionally, it incorporates Optical Character Recognition (OCR) for visible text and utilizes Whisper for searching spoken dialogue.
Users can save queries as tabs for easy access, enhancing the overall search experience. This approach not only streamlines media organization but also empowers users to interact with their content in a more intuitive manner.