Stella is a local natural language search engine designed for macOS users who need to find files by their content rather than their filenames. It belongs to the category of AI-powered desktop search tools, but with a strict focus on privacy and offline capability. Anyone who manages a large collection of PDFs, notes, or spreadsheets and struggles to locate documents by remembering exact names will find Stella invaluable. The core value is transforming file retrieval from a frustrating hunt into an intuitive conversation: you describe what you’re looking for, and Stella understands the meaning behind your query. This eliminates the need for manual organization or rigid naming conventions, making Stella a true productivity enhancer.
The primary pain point Stella solves is the time wasted searching for files when you can recall the content but not the filename. Traditional search tools like Spotlight rely on metadata, meaning you must know the exact title or extension to find a document. This breaks down when files have cryptic names, multiple versions, or when you are looking for a concept spanning several documents. For knowledge workers, researchers, and students, this friction disrupts workflow and reduces efficiency. Moreover, many cloud-based search solutions require uploading files, which compromises privacy. Stella addresses both issues: it searches by meaning and keeps everything on your machine, ensuring sensitive data never leaves your device.
The first major feature is “Meanings, Not Filenames.” Stella uses on-device AI to index the actual content of your documents—PDFs, notes, spreadsheets—and understands what they are about. Instead of matching keywords, it performs semantic search, so a query like “Q3 revenue report” finds the correct file even if it is named finance_2024_v3.xlsx. This works because Stella creates a vector representation of each document’s content, allowing it to match concepts rather than exact strings. The benefit is dramatic: you no longer need to remember how you named a file or dig through folders. You simply describe what you need, and Stella surfaces the most relevant documents instantly.
The second major feature is “No Cloud. No Trace.” Stella runs entirely on your machine, with no telemetry, no accounts, and no data leaving your device—not even temporarily. This is a completely private search engine that respects your digital sovereignty. The indexing and AI models are embedded in the 1.5 GB download, so everything happens locally. This design ensures that even highly confidential files stay secure. For users who work with sensitive information—legal documents, medical records, business plans—Stella offers peace of mind that no third party can access or analyze their data. It also works offline, on a plane or off-grid, making your library always searchable.
admin
The third feature group covers power and freedom under “What Spotlight Wishes It Was” and “Free. No Strings.” Spotlight searches filenames, but Stella reads your documents and understands their meaning, covering PDFs, notes, and spreadsheets alike. It is optimized for Apple Silicon, leveraging the M1 or later chip for fast on-device inference. Additionally, Stella is completely free with no subscription, no freemium tiers, and no account required. It works offline, so you can search even without internet access. This combination of deep semantic understanding and zero cost makes Stella a genuine upgrade for anyone relying on file search daily, without the typical freemium limitations or data collection.
The overall workflow is straightforward and designed for minimal setup. First, you select which folders Stella should watch and index—choose your documents, notes, and spreadsheets directories. Second, Stella indexes those folders locally, processing the text content of each file using on-device AI models. This indexing happens completely on your machine, taking advantage of Apple Silicon’s neural engine. Third, you search by meaning: simply describe what you are looking for in natural language, and Stella finds the right file. The interface is minimal and fast, returning results ranked by semantic relevance. There is no learning curve—you just type what you think about, and Stella handles the rest.
Concrete use cases highlight Stella’s practical benefits. A finance analyst searching for “Q3 revenue report” can locate a spreadsheet buried in a folder with dozens of similarly named files, saving minutes per search. A student preparing a research paper can find relevant PDFs by describing the concept they recall, even if the file names are random. A lawyer can quickly retrieve a specific clause from a contract by typing a paraphrase of the clause. A project manager on a flight can look up meeting notes without internet access, because Stella works offline. In each scenario, the outcome is the same: faster file retrieval with zero privacy compromise, converting a previously cumbersome task into an effortless interaction.
Stella is built for macOS users running version 13 Ventura or later, and it requires Apple Silicon (M1 or higher). The download is 1.5 GB because it includes the on-device AI models. Target users include knowledge workers, researchers, writers, students, and anyone who manages a large local file library and values privacy. Stella is free to use with no account needed, and it is currently available as a v1.0.0-beta. Windows support is coming soon. By combining local natural language search with absolute privacy and zero cost, Stella transforms how you interact with your local data—turning finding into knowing, without ever sending your files to the cloud.
Stella is for macOS users on Ventura or later with Apple Silicon Macs, including knowledge workers, researchers, students, writers, lawyers, and analysts who manage large collections of local files. It especially appeals to privacy-conscious individuals who need a free, offline-capable search tool that understands document content rather than filenames, unlike Spotlight or cloud-based alternatives.