back to top
HomeUncategorizedOsaurus: Open-Source macOS AI App for Running Local LLMs Offline

Osaurus: Open-Source macOS AI App for Running Local LLMs Offline

Osaurus Open-Source macOS AI App for Running Local LLMs Offline
Firethering Team
By Firethering Team
0
Last updated:
- Advertisement -

File Info

FileDetails
NameOsaurus
Versionv0.18.19
TypeLocal AI Harness / Offline AI Agent Platform
DeveloperOsaurus, Inc.
Size40MB
LicenseMIT License (Open Source)
PlatformsmacOS (Apple Silicon)
File Formats.dmg
AI ProvidersOpenAI • Anthropic • Ollama • Gemini • xAI • OpenRouter • LM Studio
Github Repositorygithub/osaurus
Official Sitehttps://osaurus.ai/

Description

Osaurus is a macOS-native AI harness designed around an idea “Your AI should belong to you.”

Instead of locking users into a single AI provider or cloud platform, Osaurus acts as a local control layer that sits between your AI models, tools, memory, and workflows. You can switch between local models running directly on Apple Silicon or connect cloud providers like OpenAI and Anthropic whenever you need extra power.

It is written entirely in Swift for Apple Silicon. The app feels closer to a native macOS utility. It focuses heavily on long-term context and ownership. Agents can remember conversations, manage tools, access files, execute tasks, and even run inside isolated Linux sandboxes while keeping your data on your own machine whenever possible.

Use Cases

  • Local AI assistant on macOS
  • Offline AI workflows with Apple Silicon
  • AI coding assistant with sandboxed execution
  • Long-term memory AI agents
  • Personal knowledge and file assistant
  • MCP server for AI tools and workflows
  • Voice-enabled local AI setup
  • Multi-model AI management

Screenshots

macOS AI App for Running Local LLMs Offline
Osaurus: Open-Source macOS AI App

Features of Osaurus

FeatureDescription
Fully Offline SupportRun local AI models without internet access
Native Swift AppBuilt entirely in Swift for Apple Silicon
Multi-Model SupportSwitch between local and cloud AI providers
AI AgentsPersistent agents with their own memory and tools
Long-Term MemoryIdentity, episodic memory, and pinned knowledge layers
Sandbox VMRun code safely inside isolated Linux containers
MCP Server + ClientWorks with Model Context Protocol tools
Plugin SystemExtend functionality with native plugins
Voice FeaturesOn-device transcription and voice workflows
Local Model RuntimeMLX-optimized inference for Apple Silicon
Identity SystemCryptographic identity and access keys
Relay SystemExpose agents securely over the internet
AutomationScheduled tasks and workflow watchers
Developer ToolsMonitoring, debugging, and plugin tooling
Open SourceMIT licensed project

Supported AI Providers

CategorySupported Providers
Local ModelsGemma • Qwen • GPT-OSS • Llama • DeepSeek • Apple Foundation Models • Liquid AI LFM
Cloud ProvidersOpenAI • Anthropic • Gemini • xAI / Grok • Venice AI • OpenRouter • Ollama • LM Studio

Native Plugins Included

Plugin / ToolPurpose
MailEmail access and automation
CalendarScheduling and calendar workflows
BrowserWeb browsing and automation
GitRepository and version control actions
FilesystemLocal file access and management
SearchLocal and web search capabilities
FetchData retrieval and API requests
VisionImage and visual understanding tools
MusicMusic-related integrations and controls
XLSXSpreadsheet handling
PPTXPowerPoint document support
macOS Automation ToolsNative macOS workflow automation

System Requirements

ComponentRequirement
Operating SystemmacOS 15.5+
ProcessorApple Silicon
RAM16 GB minimum
Recommended RAM64 GB+ for larger local models
OptionalmacOS Tahoe for sandbox VM support

Installation

  1. Download the .dmg file
  2. Open the installer
  3. Drag Osaurus into the Applications folder
  4. Launch the app from Spotlight
  5. Configure local or cloud AI providers
You May Like: Posturr for macOS: Smart, Privacy-First Posture Detection with Apple Vision

Download Osaurus native macOS harness for AI agent

Download Osaurus

More Than Just a Local LLM App

Osaurus is more like a personal AI operating layer for macOS.

Install it once, connect the models and tools you want, and over time it becomes a more context-aware assistant that can remember tasks, access workflows, automate actions, and run securely on your own hardware.

For users interested in privacy-first AI, local agents, Apple Silicon inference, or building a long-term AI workspace on macOS, Osaurus is one of the more ambitious open-source projects currently emerging in the local AI space.

Want more stories worth your time?

Add us to your Google favorites. We cover the tech stories, AI developments, and open-source projects that are easy to miss in the noise.

Add as a preferred source on Google

Don’t miss any Tech Story

Subscribe To Firethering NewsLetter

You Can Unsubscribe Anytime! Read more in our privacy policy

LEAVE A REPLY

Please enter your comment!
Please enter your name here

YOU MAY ALSO LIKE
LLM Wiki Desktop

LLM Wiki: An AI Knowledge Base That Builds Itself

Firethering Team - 0
Most AI tools can answer questions about your documents. But there's a problem with the usual approach. You add a bunch of files, ask a question and the AI searches through them to find an answer. Ask something else later and it does much of the same work again. LLM Wiki on the other hand, uses an LLM to build a persistent, structured wiki from them. Add a collection of documents and LLM Wiki analyzes them, creates pages for important entities and concepts, connects related information and keeps the knowledge base updated as you add more sources. That means your knowledge doesn't just sit inside a folder waiting to be searched. It gradually becomes an interconnected collection of information that you can browse, search and ask questions about.
hister search engine

Hister: Your Own Private Search Engine for Web Pages and Files

Firethering Team - 0
You know that page you read three months ago and somehow can’t find again? Hister is built for exactly that problem. It turns the web pages you visit and the files you keep into your own searchable index, so you can search the actual content instead of trying to remember a title, URL, or where you saved it.
Atomic Chat App

Atomic Chat: Run Open-Weight LLMs Locally on Windows, macOS & Linux

Firethering Team - 0
Want to run an AI model locally, but still use it with the tools you already have? Atomic Chat makes that possible. It lets you run open-weight LLMs from Hugging Face on your own computer, then exposes them through an OpenAI-compatible API so coding agents, CLIs, IDE plugins and other apps can use your local models too. You can run models such as Llama, Gemma, Qwen, Mistral and Phi, use Atomic Chat as a regular AI chat app, or connect it to tools such as OpenCode, Goose and Kilo Code. Your local conversations and API keys can stay on your machine, while cloud providers such as OpenAI, Anthropic, Mistral and Groq are available when you need them. Under the hood, it also supports multiple inference engines and performance features such as speculative decoding, Flash Attention and TurboQuant on supported models and hardware. So you're getting more than a local chatbot. Atomic Chat can act as the local AI layer behind the rest of your setup.