A native macOS app for chatting with local LLMs served by Ollama, with multiple threads, markdown rendering and vision image support.
Jeidoban/Ironsmith
A free macOS menu bar app that generates, builds, and saves small native SwiftUI Mac apps from text prompts using AI.
- Stars
- 344
- Forks
- 29
- Open issues
- 0
- Last commit
- 11 days ago
What is it?
Ironsmith lets you describe a small utility in plain language and generates a native SwiftUI app that you can run, edit, and export to your Applications folder. Each generated app is a Swift package built with the Xcode Command Line Tools, so full Xcode is not needed, and the app itself is signed with sandboxing and hardened runtime enabled by default.
The app works with local models through Ollama and any OpenAI-compatible API such as LM Studio or Llama.cpp, and also supports hosted models from OpenAI, Anthropic, and Gemini using your own API keys. Users can pick Ironsmith's built-in agents for small apps or OpenAI's Codex for larger projects.
The project also includes the Ironsmith Store, where published apps can be searched by capability and used as starting points for remixing. It requires macOS 26 or newer and supports both Intel and Apple Silicon Macs.
Key features
- 01Generates native SwiftUI Mac apps from a text prompt.
- 02Runs with local models via Ollama or any OpenAI-compatible API.
- 03Supports hosted models from OpenAI, Anthropic, and Gemini with your own keys.
- 04Builds apps with Xcode Command Line Tools, without requiring full Xcode.
- 05Sandboxes generated apps by default, with sensitive permissions opt-in.
- 06Lets you remix apps from the Ironsmith Store as starting points.
Use cases
- Building a one-off utility, such as a file renamer or PDF splitter, without learning Swift.
- Prototyping small macOS tools while keeping data on a local model.
- Starting from an existing Store app and adapting it to a specific workflow.
Quick start
- 1
Download the latest Ironsmith build from GitHub Releases or the website
- 2
Install the Xcode Command Line Tools manually (optional, macOS prompts if missing)
xcode-select --install - 3
Build and run the development app from source
script/build.sh run
Things to know
- Requires macOS 26 or newer, so it does not run on older Macs or other platforms.
- Apple Intelligence must be enabled for app icon generation and the built-in Foundation Model.
- Hosted models need your own API keys or an Ironsmith sign-in, while local models need a separate setup such as Ollama.
- Licensed under GPL-3.0, which requires derivative works to be distributed under the same license.
In collections
Topics
FAQ
- What is Ironsmith?
- A free macOS menu bar app that generates, builds, and saves small native SwiftUI Mac apps from text prompts using AI. Ironsmith lets you describe a small utility in plain language and generates a native SwiftUI app that you can run, edit, and export to your Applications folder. Each generated app is a Swift package built with the Xcode Command Line Tools, so full Xcode is not needed, and the app itself is signed with sandboxing and hardened runtime enabled by default.
- Is Ironsmith free to use? What license does it have?
- Ironsmith is open source under the GPL-3.0 license. Check the license terms before using it in commercial products.
- Is Ironsmith still maintained?
- Actively maintained. The last commit to Ironsmith was on Oct 28, 2026.
- What can I use Ironsmith for?
- Building a one-off utility, such as a file renamer or PDF splitter, without learning Swift. Prototyping small macOS tools while keeping data on a local model. Starting from an existing Store app and adapting it to a specific workflow.
- How do I install Ironsmith?
- Download the latest Ironsmith build from GitHub Releases or the website Install the Xcode Command Line Tools manually (optional, macOS prompts if missing): xcode-select --install Build and run the development app from source: script/build.sh run
- Where can I get Ironsmith?
- The source code and installation instructions are on GitHub: https://github.com/Jeidoban/Ironsmith
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A fully local, voice-enabled AI assistant that browses the web, writes code and plans tasks on your own hardware, as an alternative to Manus AI.
A native macOS AI workspace in SwiftUI that runs locally, with chat, agent tools, email, calendar, memory and document editing.
A macOS meeting note-taker that transcribes both sides of a call locally and surfaces talking points from your own notes.
Open-source voice dictation app for Mac: hold Fn, speak, and cleaned-up text is typed wherever your cursor is.
Open-source AI developer workspace that builds full-stack web apps in the browser, letting you pick any supported LLM per prompt.