Dotient is a $5 one-time desktop app that uses on-device AI to search your files by describing what they look like — not by guessing filenames. It runs fully offline, keeps everything private, and just hit #4 on Product Hunt.
See it in action in the official launch video:
Meet Dotient — Official launch video showcasing the visual search and graph view features. Video: dotient/YouTube.
What Is Dotient?
Dotient is a local-first desktop application built by solo developer Declan (@localdeclan) that brings semantic and visual search to personal files. Built with Tauri 2 (Rust backend, Svelte frontend), it runs a quantized vision ML model entirely on your machine — no data ever leaves your device.
Launched June 28, 2026, Dotient reached #4 on Product Hunt with 184 upvotes and 217 followers. At $5 one-time for up to three computers, it’s a refreshingly rare pricing model in a market dominated by $10–20/month subscriptions.
If you’ve ever stared at Windows File Explorer knowing a file exists somewhere in a folder structure you built years ago, Dotient speaks to that frustration. The Product Hunt reception confirmed it: people are tired of filename-based search and want a smarter, private way to find their own content.
Visual Search: Describe It, Find It
Dotient’s core feature is visual search. Instead of typing a filename, date, or extension, you describe what you remember: “beach sunset from last summer,” “bird with yellow beak on a feeder,” “screenshot of the Q3 budget.” The on-device vision model embeds every file’s visual content and matches it against your natural-language query.
The search engine uses a hybrid BM25 + semantic vector approach — keyword and semantic search run in parallel so neither method misses what the other would catch. This is a smart engineering tradeoff that addresses the biggest trust issue with AI-powered search: “Why didn’t this result show up?”
For PDFs, Dotient indexes both the text layer and embedded images. A rectangle selection tool lets you search specific document regions — useful for diagrams, infographics, or tables. Full OCR isn’t here yet (the maker has it on the roadmap), but the vision model already understands text appearing inside images to a meaningful degree.
Graph View, Cubbies, and Signals — More Than a Search Engine
Dotient pairs its search with three tools that turn file discovery into an active, visual experience:
The Graph View
Every file in your archive appears as a node in an interactive network visualization. Edges form automatically between files based on ML similarity scores. The app detects clusters of related content, and you can label and explore your file structure as a living map. It’s a genuinely new way to browse a digital archive — one that reveals connections you didn’t know existed.
Cubbies (Collections)
Cubbies are user-defined collections — mood boards, project folders, themed sets — that complement algorithmic search with manual organization. Each cubby supports annotations, to-do lists, and freeform notes. They bridge the gap between “the AI found it” and “I need to organize it.”
Signal Tuning
The Signal system is Dotient’s answer to improving search over time. When results miss something, you flag the right file as relevant. The system shifts its embedding space based on your feedback, creating personalized filters that improve with each interaction — a feature most cloud-based competitors don’t offer at any price.
The $5 Anti-Subscription Pitch
In a market where most AI-enhanced software charges $10–20 per month, Dotient’s $5 one-time fee (covering up to three computers with all features and future updates included) feels like a statement. The app has no tiers, no feature gates, no upsells. You buy it once and it’s yours.
The only time Dotient touches the internet is to verify your license key. Everything else — ML models, embeddings, search index, database — lives locally at C:\Users\[username]\AppData\Roaming\com.dotient.app. A detailed breakdown of the tech stack is available on HuntScreens, which confirms the Tauri 2 + Svelte + Rust architecture and the on-device quantized model approach.
The maker has been transparent about tradeoffs. Model versioning, for example, is acknowledged: when the embedding model updates, older indexed files may not align perfectly. The plan is lazy migration — keep the old index queryable, slowly re-embed with the new model, dual-query merge until the old index drains. Explore AI Tools’ 7/10 editorial review highlights this as a realistic approach to a genuinely hard problem in local-ML software.
How Dotient Compares
Dotient enters a space with several established players. Here’s how it stacks up against the alternatives:
| Tool | Approach | Key Difference |
|---|---|---|
| Windows File Explorer | Index-based, filename/metadata | Dotient uses visual semantics, not filenames |
| Everything (voidtools) | Lightning NTFS filename search | No content or visual search |
| Eagle | Paid organizer with tagging | Dotient is $5 one-time vs Eagle’s subscription |
| Google Photos | Cloud-based semantic search | Dotient is local, no upload, no subscription |
The closest parallel in recent TekMag coverage is FlowLauncher, another desktop utility that reimagines a core Windows experience. Where FlowLauncher transforms the keystroke launcher, Dotient transforms file search — both proving the Windows desktop is still ripe for thoughtful reinvention.
Who Is Dotient For?
Dotient targets three overlapping user groups:
- Photographers and videographers with thousands of shoots across hundreds of folders. Describe the shot and find it instantly — no tagging, no metadata review, no manual sorting.
- Designers and art directors drowning in reference images, mood boards, and inspiration dumps. Use Cubbies to group, annotate, and organize.
- Anyone with an unorganized personal archive — years of photos, scans, screenshots, and document dumps that are functionally lost in folder hell.
If your File Explorer is a graveyard of unsorted Downloads folders, Dotient offers a way out that doesn’t require hours of manual cleanup.
Final Verdict
Dotient is a well-executed tool for a specific pain point. If you have a tidy folder structure and remember every filename, you’re not the target. But for the growing number of people whose digital lives outpace their organizational systems, Dotient’s on-device AI, one-time pricing, and privacy-first architecture make it one of the most practical local-search tools available today.
At $5, it’s low-risk enough to try — and if it saves you one frustrated “where is that file” moment, it’s already paid for itself.
Dotient launched in a week packed with major news. We covered it alongside other top developments in our weekly AI roundup. And if Windows performance is on your mind, our breakdown of Windows 11’s June 2026 Low Latency Profile explains why this has been a big month for Windows improvements.
Frequently Asked Questions
Does Dotient work fully offline?
Yes. Once installed and indexed, everything runs locally. The only internet connection is a single call to verify your license key. Your files, the ML model, embeddings, and search index all stay on your machine.
What file types can I search with Dotient?
Dotient primarily handles visual media (images, photos, screenshots) and PDFs (text layer and embedded images). The vision model also understands text inside images to some extent. Full OCR for scanned documents is on the roadmap.
Is my data safe with Dotient’s AI?
Absolutely. The ML model runs entirely on your device — nothing is sent to a server. The maker explicitly states: “Absolutely nothing leaves your device. No server besides the small API that simply checks if your license key is valid.”
Originally published on TekMag
Top comments (0)