Osaurus, the Native macOS Harness for Local AI Models, Celebrates 7.3K GitHub Stars and #2 Product of the Day on Product Hunt

Open-source macOS app for running local models and cloud AI agents on the Mac has passed over 185K downloads

San Francisco, CA, July 20, 2026 (GLOBE NEWSWIRE) -- Osaurus, the native macOS harness built around local models, has crossed 7.3K stars on GitHub and over 185K downloads, and finished as the #2 Product of the Day on Product Hunt following its latest launch. The open-source project, built entirely in Swift, lets people run local models directly on their own Macs, online or fully offline, without sending their data to an outside server.

Osaurus, the native macOS harness for running AI agents locally on the Mac.

Osaurus works as a control layer, or harness, built local-model-first: it runs open models directly on the Mac's own silicon and only reaches hosted cloud models through the same interface when a task calls for them. Every agent runs inside a hardware sandbox that keeps memory, files, and tools under the user's control, and the app functions as a full Model Context Protocol (MCP) server with more than 20 native plugins spanning mail, calendar, browser, vision, version control, and the filesystem - all able to run against a local model with no cloud round trip. Osaurus also uses custom quantization, which lets people run larger models at smaller quants than they could normally run on their hardware. It can also share model directories with LM Studio, Ollama, and MLX, so users can try it out without having to redownload models they already have.

Despite that technical depth, Osaurus is built to be approachable: there is no coding, no configuration files, and no command line required, so non-technical professionals can download the app, connect a model, and start using AI agents within minutes.

The milestones stand out because Osaurus reached them without paid marketing. The project has grown by word of mouth since its first release, adding persistent memory, autonomous execution, cryptographic identity, and voice control as its community asked for them - every one of them running on local models by default. What started as a lightweight way to run local models has become a full agent platform shaped largely by contributor and user feedback.

"People want the power of modern AI without handing over their files, their memory, and their identity to someone else's cloud," said Terence Pae, co-founder of Osaurus. "Local models are the foundation of Osaurus, not an afterthought - everything runs on the user's own machine by default, so an agent can browse files, use the browser, and automate system sensitive context that never leaves the Mac. The community built this alongside us, and every milestone belongs to them as much as to us. You don't need to be an engineer to use it - if you can install an app, you can run your own local AI agents."

The release arrives as interest in local models and privacy-first AI continues to climb, with users looking for alternatives to cloud-only assistants that require every request to leave the device. By keeping local models, memory, and tools on the Mac itself, Osaurus aims to give individuals and teams the benefits of autonomous agents while leaving them in control of their own data.

Osaurus is free to download and MIT-licensed, and it runs local models on modern Mac hardware. More information, documentation, and the full source code are available at osaurus.ai.

About Osaurus

Osaurus is the native macOS harness for AI agents - any model, persistent memory, autonomous execution, and cryptographic identity, built in Swift and able to run fully offline on Apple Silicon. The open-source, MIT-licensed platform lets Mac users run local and cloud AI models side by side inside a hardware sandbox, acts as a full Model Context Protocol (MCP) server, and ships with 20+ native plugins spanning Mail, Calendar, Browser, Vision, Git, and the filesystem. Free to download, Osaurus has grown by word of mouth to 7,300+ GitHub stars and 185,000+ downloads.

Press Inquiries

Annie
annie [at] osaurus.ai
https://osaurus.ai

Legal Disclaimer:

EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.