Key Takeaways
- Hyperlink (hyperlink.nexa.ai) indexes PDFs, notes, screenshots, and slides on your own computer and answers questions with citations to the source file.
- All processing happens on-device — no file content leaves your machine, and it works offline once files are indexed.
- Runs on Apple Silicon Macs and Windows PCs with an NVIDIA GPU, AMD GPU, or Snapdragon X (ARM) chip; Intel integrated-graphics support is listed as coming.
- Closed-source (proprietary) — there is no public GitHub repository to inspect.
- Developed by Nexa AI, which became a Qualcomm company in March 2026; Hyperlink itself remains a separately branded, actively distributed product.
- NVIDIA's own blog post on RTX acceleration for Hyperlink reports indexing up to 3x faster and inference up to 2x faster on RTX GPUs versus CPU-only — a vendor claim this review has not independently reproduced.
📍 In One Sentence
Hyperlink is a free, closed-source desktop app from Nexa AI that indexes your local files and answers natural-language questions about them with citations, processing everything on-device with no cloud upload.
💬 In Plain Terms
Think of it as a private, offline search engine for your own PDFs, notes, screenshots, and slides — you ask a question in plain English, and it points you to the exact file and passage the answer came from, without sending any of that content to a server.
📌Note: This review is the deep-dive companion to Hyperlink's entry in the Local LLM Software Directory — see that page for how Hyperlink compares at a glance to dozens of other local AI tools.
📌Note: Hyperlink is closed-source, so this review cannot independently verify its on-device/no-upload claims the way it could for an open-source tool. Those claims come from Nexa AI's own documentation and are corroborated by NVIDIA's official blog post on the RTX integration (see Sources) — not from PromptQuorum's own network testing.
What Is Hyperlink?
Hyperlink is a desktop application that turns the files already on your computer into a searchable, question-answerable knowledge base — without uploading any of them anywhere. It indexes PDFs, notes, screenshots, and presentation slides, then lets you type a question in plain language and get back a direct answer with a citation pointing to the specific file (and, where applicable, passage) the answer came from.
Nexa AI positions Hyperlink as an on-device alternative to asking a cloud AI search tool to search the open web — instead, the search space is the set of files you point it at on your own drive, and the entire indexing and question-answering pipeline runs locally using Nexa AI's own inference stack plus open models from Hugging Face.
The product has no public source code repository, which distinguishes it from most tools reviewed in this directory — it is closed-source, proprietary software, distributed as a free download rather than as an open project a reader can audit or self-build.
Nexa AI and the Qualcomm Acquisition
Hyperlink is developed by Nexa AI, a company that became a Qualcomm company in March 2026. Before the acquisition, Nexa AI built on-device AI infrastructure including the NexaSDK toolkit and small-model tooling for running AI locally on phones, laptops, and other edge devices.
After the acquisition, Nexa AI and Qualcomm continued to publish joint developer tooling — including GenieX, an open-source runtime for running generative AI models on Qualcomm chipsets — while Hyperlink has continued to be distributed and updated as a separately branded consumer product.
This review treats Hyperlink as the reviewable product and Nexa AI/Qualcomm as its developer; readers evaluating Hyperlink for long-term use should check Nexa AI's own announcements for how the acquisition affects the app's roadmap or support going forward, since PromptQuorum has not independently verified any post-acquisition product commitments beyond what Nexa AI and Qualcomm have publicly stated.
What Can You Do With Hyperlink?
Hyperlink's feature set centers on turning a folder of files into something you can ask questions of directly, per Nexa AI's own product pages and its NVIDIA and Qualcomm partner announcements.
- Natural-language file search — ask a question in plain English (or another supported language) instead of typing exact keywords; Hyperlink's indexing is built to understand query intent, not just match text.
- Cited answers — responses point back to the specific file (and, where supported, the passage within it) the answer was drawn from, so you can verify the source rather than trusting an unattributed summary.
- Broad file-type support — PDFs, plain notes, screenshots, and presentation slides are indexed; image content is processed alongside text-based documents.
- Fully offline operation — once files are indexed, Hyperlink answers questions without needing an internet connection, since no query or file content is sent to a remote server.
- Model choice — Hyperlink runs on Nexa AI's own inference engine and supports a range of open models sourced from Hugging Face, letting you trade off between faster, lighter models and larger, more capable ones.
- Hardware-accelerated indexing and inference — on supported NVIDIA RTX GPUs, Nexa AI and NVIDIA report faster indexing and inference versus CPU-only processing (see the hardware section below for the specific figures and their source).
Usage Examples: Three Ways to Use Hyperlink
These are workflows Nexa AI itself documents as intended use cases, not hypothetical scenarios.
Get Hyperlink for macOS and Windows
Hyperlink is available as a free download for macOS and Windows, distributed from Nexa AI's own site and listed on the Microsoft Store and Product Hunt. Links below always verify against those pages directly, since download URLs and build requirements can change between releases.
macOS (Apple Silicon)
- Get It:
- Download from hyperlink.nexa.ai
Windows (NVIDIA/AMD GPU or Snapdragon X)
- Get It:
- Download from hyperlink.nexa.ai or via the Microsoft Store
Product listing
Intel integrated-graphics support was listed as "coming soon" as of this review's publish date — check the official download page for current platform coverage before assuming your specific Windows configuration is supported.
Hyperlink Pricing and License
Hyperlink is distributed as a free download, and third-party app directories describe it as free for both personal and workspace use. This review found no confirmed paid tier, subscription, or enterprise pricing page during research — that absence is not the same as a guarantee no paid tier exists, so verify current pricing directly on hyperlink.nexa.ai before making a decision based on cost.
Price
- Status:
- Free to download and use, per Nexa AI's own listing and third-party app directories; no confirmed paid tier found
Source code
- Status:
- Closed-source — no public GitHub repository; this review classifies the license as proprietary
Account requirement
- Status:
- Not independently confirmed either way in this review — check the app's own onboarding flow
Because Hyperlink is closed-source, PromptQuorum cannot point to a public LICENSE file the way it can for an open-source tool. Treat "proprietary" here as the classification, not a specific license text, and confirm current terms directly with Nexa AI before any commercial or workspace deployment.
System Requirements
Hyperlink's supported hardware spans Apple Silicon Macs and several Windows GPU/chip configurations. Requirements below are drawn from Nexa AI's own platform-support statements and third-party app-directory summaries; confirm exact minimums on the official site for your specific machine.
Apple Silicon Mac
- Support Status:
- Fully supported
- Notes:
- No separately reported RAM minimum beyond general macOS requirements
Windows with NVIDIA or AMD GPU
- Support Status:
- Fully supported
- Notes:
- NVIDIA RTX GPUs get additional acceleration — see the note below
Windows on Snapdragon X (ARM)
- Support Status:
- Fully supported
- Notes:
- Runs via Qualcomm Hexagon NPU-class hardware, consistent with Nexa AI's on-device AI heritage
Windows, Intel integrated graphics only
- Support Status:
- Listed as coming soon; reported to need around 32 GB RAM
- Notes:
- Verify current status before purchasing hardware to run Hyperlink
On NVIDIA RTX-accelerated hardware, NVIDIA's own blog post announcing the integration reports indexing up to 3x faster and inference accelerated up to 2x versus CPU-only processing, with an example of a 1 GB folder's index time dropping from roughly 15 minutes to 4-5 minutes. These are vendor-published figures from NVIDIA and Nexa AI, not benchmarks PromptQuorum ran independently — treat them as directional rather than guaranteed on every machine.
Who Should Use Hyperlink?
Whether Hyperlink fits depends on whether you want a ready-made, closed-source desktop app rather than an open-source tool you can self-host or inspect.
Competitors and Alternatives
Hyperlink sits in the local document-search and personal-knowledge-assistant category — tools that let you ask questions of your own files or notes with a local or self-hosted backend. Here is how it compares to other options reviewed on PromptQuorum; see the Local LLM Software Directory for the full catalog.
- Khoj — an open-source, self-hostable "AI second brain" with very similar positioning to Hyperlink (search and chat with your own documents and notes), but open-source rather than closed. See the Khoj review.
- PrivateGPT — an open-source, self-hosted tool for asking questions of your own documents entirely offline, aimed more at developers comfortable running a local server than a polished consumer app. See the PrivateGPT review.
- Quivr — an open-source, self-hosted "second brain" for uploading and querying your own files and knowledge base. See the Quivr review.
- AnythingLLM — an open-source desktop and self-hosted app for chatting with your own documents, with a more configurable backend (choice of local or cloud models) than Hyperlink's fixed pipeline. See the AnythingLLM review.
Common Mistakes When Evaluating Hyperlink
Most confusion about Hyperlink comes from assuming it works like an open-source tool, or overreading vendor performance claims as independently verified benchmarks.
Frequently Asked Questions
What is Hyperlink?
Hyperlink (hyperlink.nexa.ai) is a free desktop app from Nexa AI that indexes local files (PDFs, notes, screenshots, slides) and answers natural-language questions about them with citations to the source file, processing everything on-device.
Is Hyperlink free?
Yes, Hyperlink is free to download and use, per Nexa AI's own listing and third-party app directories. This review found no confirmed paid tier — verify current pricing directly on the official site before assuming that will never change.
Is Hyperlink open source?
No. Hyperlink is closed-source, proprietary software with no public GitHub repository, unlike most tools reviewed in this directory.
What platforms does Hyperlink support?
Apple Silicon Macs, and Windows PCs with an NVIDIA GPU, AMD GPU, or Snapdragon X (ARM) processor. Intel integrated-graphics-only Windows support was listed as coming soon as of this review's publish date.
Does Hyperlink upload my files to the cloud?
No, according to Nexa AI's own documentation: Hyperlink's processing is designed to run on-device, with no file content sent to a remote server, and the app works offline once files are indexed. This review has not independently verified that claim, since Hyperlink is closed-source.
Who develops Hyperlink?
Nexa AI, a company that became a Qualcomm company in March 2026. Hyperlink continues to be distributed and updated as a separately branded consumer product.
How much faster is Hyperlink on an NVIDIA RTX GPU?
NVIDIA's own blog post on the integration reports indexing up to 3x faster and inference up to 2x faster on RTX-accelerated hardware versus CPU-only, with a 1 GB folder example dropping from about 15 minutes to 4-5 minutes. This is a vendor-published figure, not an independent PromptQuorum benchmark.
What file types does Hyperlink support?
PDFs, plain notes, screenshots, and presentation slides, per Nexa AI's own product documentation and partner announcements.
What is the closest open-source alternative to Hyperlink?
Khoj is the closest match in positioning — both are described as an "AI second brain" for searching and chatting with your own documents — but Khoj is open-source and self-hostable, while Hyperlink is closed-source. See the Khoj review for details.
