DigUp Launched Semantic Local File Search for Mac
The open-source tool enables local file retrieval on Apple Silicon Macs using a 740-million-parameter embedding model.
Updated on Oct. 14, 2026 in Artificial Intelligence
Live Poll
Do you trust local AI file search tools more than cloud-based alternatives?
DigUp has released a free, open-source application that allows Apple Silicon Mac users to perform semantic searches across local documents, photos, and media. The software runs entirely on-device, bypassing cloud storage by utilizing Google's EmbeddingGemma 2 model.
Why it matters
This development shifts file management from rigid, name-based hierarchies to descriptive, intent-based retrieval. By running the search index locally, it provides a privacy-focused alternative to cloud-reliant search solutions.
The application relies on the 740-million-parameter EmbeddingGemma 2 model, which requires an 865MB download. This model enables the mapping of text, images, and audio into a shared semantic space for matching.
The players
DigUp
An open-source software project that provides local, semantic file search capabilities for macOS users.
The technology company behind the EmbeddingGemma 2 model used for local vector-based indexing.
The details
DigUp functions by having users select folders for local indexing, which the app then converts into vector representations—a numerical method for mapping data to show relationships between concepts. By using a shared embedding space, the software matches queries to the semantic meaning of photos, PDFs, audio, and video content. The tool also includes a specialized feature for searching code based on function descriptions rather than syntax.
Timeline
October 14, 2026: DigUp became available for users on Apple Silicon Macs.
The Tech Race
DigUp marks a departure from traditional keyword-based desktop search engines like Apple's Spotlight by prioritizing semantic meaning over file-name metadata. It positions itself against broader industry efforts to move AI inference and vector indexing from the cloud to the edge.
Users require an Apple Silicon Mac running macOS 14 or later to operate the software. Once installed, the application allows for content-based retrieval across local file directories, effectively replacing traditional text-string searches for documents, media, and code.
The takeaway
DigUp brings semantic search capabilities to the local desktop environment, enabling users to find files based on context rather than explicit naming. Interested users should monitor the project's open-source repository for future updates to the EmbeddingGemma 2 integration or model compatibility.
Further reading
For more on the development of edge-based machine learning tools, see the latest updates in Artificial Intelligence.
Source note: This article includes information reported by Digital Trends.
Live Poll
Do you trust local AI file search tools more than cloud-based alternatives?
