workflowstacks
mcp-server

osaurus: Run Offline AI Agents on macOS with Persistent Memory

osaurus lets founders deploy any LLM locally on macOS with autonomous execution, cryptographic identity and Swift‑native performance. 8.0k+ GitHub stars.

From the repo · “Own your AI. The native macOS harness for AI agents -- any model, persistent memory, autonomous execution, cryptographic identity. Built in Swift. Fully offline. Open source.”

advanced⏱ 1-2 hours💵 Free (self-hosted)
8,066 stars+18★ this week483 forksSwiftGuide quality 8/10 · reviewed Oct 8Updated 10/22/2026100% free · open source
What it is

Run AI agents on your own computer without an internet connection.

Real use case example

"A founder wants to create a welcome email sequence when a new customer signs up on their website. They first write a function to listen for new signups, then use osaurus to interact with the AI to generate a personalized email. Finally, they use osaurus to send the email to the new customer."

Best insideClaude DesktopClaude Code
When to use it
  • •You need to deploy AI models on macOS devices without relying on cloud services.
  • •Your application requires autonomous AI execution with persistent memory.
  • •You want a fully offline, open-source solution for running AI agents on macOS.
Ready-to-paste prompt
Run `osaurus start --model my_model` to start an autonomous AI model named 'my_model' with osaurus.
Heads up: Ensure you have Xcode and the Swift compiler installed on your macOS, as osaurus is built in Swift and requires these tools to build and run.
Saves to your device
Use with Claude
New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install
Claude Code
mkdir -p ~/.claude/skills/osaurus && curl -fsSL https://workflowstacks.com/api/skills/osaurus/claude-skill -o ~/.claude/skills/osaurus/SKILL.md
Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)
claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How osaurus: Run Offline AI Agents on macOS with Persistent Memory works
Codeflow
Free to inspect

Osaurus: Run Offline AI Agents on macOS with Persistent Memory is a very large Swift project (~665k lines across 1425 code files, plus 959 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, MIT license, has a test suite.

Size
Very large codebase
~665k lines · 1425 code files · days to read — use, don't read
Setup
Developer setup
A real software project. Use it via its install path; don't expect to read it all.
Runs on
Swift
No API keys detected
Swift 78%C 20%Shell 1%
Where to start reading
  1. 1
    README.md
    Start here — what it does and how to install it
  2. 2
    AGENTS.md
    The instructions the AI actually follows
  3. 3
    App/osaurus/osaurusApp.swift
    Inside App/ — the main logic begins here
What's in each folder
App/Core code — the actual logic41 files
docs/Documentation171 files
Packages/OsaurusCore/Package2226 files
Packages/OsaurusEvals/Package672 files
scripts/Helper scripts80 files
Packages/OsaurusCLI/Package76 files
reports/Folder19 files
Packages/OsaurusRepository/Package15 files
READMEHas testsDocumentedCI checksMIT licenseUpdated this month
Quick Actions
Details
Creator
osaurus-ai
Language
Swift
Category
mcp-server
Published
8/17/2025

Are you the creator of this tool? and earn 85% of every sale.

Show it off in your README:

Featured on WorkflowStacks
[![Featured on WorkflowStacks](https://workflowstacks.com/api/badge/osaurus.svg)](https://workflowstacks.com/skills/osaurus?utm_source=github&utm_medium=badge)
🔥 Hot this week, in your inbox
Skills like this, every Monday.

The five fastest-growing open-source AI skills, ranked by GitHub star growth. One email a week, unsubscribe anytime.

Do this with a working workflow

Pages we wrote and tested that put a skill like this to work — importable n8n templates, outcome guides and Claude Desktop configs.

Free n8n template

Cold Email Personalizer

Turn a list of prospects into genuinely specific, personalized cold-email openers — each scored 1-5 for specificity so you can kill the generic ones.

Outcome guide

Personalize your cold emails at scale

You keep prospects in a Google Sheet with a note about each. The workflow writes a specific one-line opener per prospect and scores it 1–5, so you can send the strong ones and fix the weak ones.

MCP config

Memory (knowledge graph) MCP config

Persistent memory across chats via a local knowledge graph.

Related skills

More mcp-server tools founders pair with this one.

mcp-server★ 355,695
Rea: Reverse Engineer Apps
Inspect apps down to binary level with Rea. For founders needing code clarity.
mcp-server★ 156,361
open-webui: Self‑hosted AI Chat UI for Faster LLM Integration
Launch a sleek, web‑based AI interface that works with Ollama, OpenAI API, and more—ideal for founders needing a quick, customizable LLM front‑end. 153k+ GitHub stars.
mcp-server★ 131,266
graphify
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.
mcp-server★ 1,908
LangAlpha
Claude Code for Financial Market
mcp-server★ 17,721
rocketride-server: Turbocharge LLM Pipelines with C++ Speed
Scale AI workflows from your IDE using 50+ extensible nodes, 13 model providers and 8 vector databases—ideal for founders building production LLM products. 18k+ GitHub stars.
mcp-server★ 29,891
FastGPT: Deploy Powerful Q&A Systems Fast with 30k+ GitHub Stars
FastGPT lets founders build and launch sophisticated question‑answering apps using built‑in data processing, RAG retrieval, and visual AI workflows—no heavy setup required.