My LLM
Polish an email draft. Turn a long passage into a short summary. Read an announcement in another language. My LLM is a small desktop app that helps with these everyday writing tasks, right alongside the apps you already use.
Save instructions for polishing, summarizing, translating, and more as tasks, then call them from the menu bar or a hotkey. Instead of writing the same instructions each time, choose the text and the task you want to run.
Download / Installation and user guide
Figure 1: The original notes are at the top, with the summary below. The bottom of the window shows the selected task, provider, and model, along with Run and Copy. You can give tasks your own names.
Your everyday text, your own tasks
Here are a few ways to use My LLM.
| What you want to do | How to set up a task |
|---|---|
| Make text easier to read | Save instructions to improve the wording while preserving its meaning and tone |
| Find the key points in a long note | Save instructions to produce bullet points or a short paragraph |
| Understand a technical term | Save instructions for a brief, plain-language explanation |
| Read text in another language | Use the translation task to detect the source language and translate into your chosen language |
| Fit your own workflow | Add a custom task with instructions for your preferred writing style or output format |
Each task can have its own instructions, display name, hotkey, and model. Start with the sample tasks, then add the ones you use most often.
Copy and call a task, or paste and run
First, install the app, prepare a provider, and choose a model for your task. Follow the setup instructions in the user guide, then use this workflow.
- Choose Open Window from the My LLM menu in the menu bar (or the system tray on Windows).
- Type or paste text into Input.
- Choose a task and click Run.
- When the result appears in Output, click Copy and paste it back into your original app.
Once you are familiar with the app, copy some text and choose a task from the menu to capture the text and run it in one action. You can also assign hotkeys to tasks and automatically copy completed results.
Results appear progressively through streaming. The floating window can stay alongside your other work.
Local and cloud models
My LLM connects to Ollama or the cloud APIs provided by OpenAI and Anthropic.
- Use Ollama models
Run local models on your computer, or use cloud models through a signed-in Ollama instance. Local use requires downloading a model and having enough memory to run it. For cloud access, see Ollama's sign-in instructions. - Use a cloud API
Send your text and instructions to the API you configure and receive the result. Check the service's API key requirements, terms, and pricing.
You can choose a provider for each task. For example, use a local Ollama model for translation and a cloud model for polishing text. Both the endpoint and the model determine where your text is processed. Selecting a cloud model through Ollama sends your text to the cloud as well.
Translation as a task
Translation detects the source language and chooses a target using your settings. For example, you can translate English into Japanese, and Japanese or any other non-English input into English.
Figure 2: An English announcement translated into Japanese. The label at the bottom shows that the task uses Ollama's translategemma:12b model.
Use the built-in instructions for TranslateGemma, or write your own translation instructions for another model. The app's interface language and the target language for translation are separate settings.
Configure tasks in a settings window
Choose Settings… from the menu to configure providers, tasks, and translation. Write your instructions, give the task a name, and assign a hotkey. The settings reference explains each field.
Figure 3: An example summary task. My LLM combines the instructions in Instruction with the text you want to process. The values shown are examples.
Download and requirements
My LLM is released under the MIT license. Downloads are currently available for macOS and Windows.
| System | Download |
|---|---|
| macOS 13 or later, Apple Silicon | File ending in darwin-arm64.dmg |
| macOS 13 or later, Intel | File ending in darwin-x86_64.dmg |
| Windows 11, x86_64 | File ending in windows-x86_64.zip |
Download from GitHub Releases. Models are provided separately from the app. To run a local model through Ollama, check Ollama's system requirements and the model's memory requirements as well.
For connection setup and your first task, continue to the My LLM user guide. The source code is also available.
Where to go next
| What you want to do | Read |
|---|---|
| Install the app and get your first result | User guide: prepare the app and provider |
| Process copied text quickly | User guide: everyday use |
| Save your own instructions as a task | User guide: create a task |
| Look up providers, models, translation, and other settings | Settings reference |
| Find help when something does not work | User guide: troubleshooting |
This introduction and manual cover the desktop GUI implemented in Rust.