Free local AI for your computer.

Yes, for personal and professional use, with no message limit. That is possible because the model runs on your computer: there are no AI servers to pay for your answers. Twoody is designed and funded in France by its creator, Jimmy Douieb. No paid offer, no advertising, and no data sold.

For everyday work, often yes: writing, summarizing, translating, questions about your own documents, for free and in private. A model that runs on a personal computer is smaller than the large online models: for the hardest questions, they remain stronger.

Yes, once the model is installed. With Wi-Fi off, Twoody answers, reads and indexes your documents, remembers what you tell it, keeps your projects and reminders, and sorts your files with “Work in your folders”. Dictation too, once its voice model is downloaded (a Mac that recognizes your language by itself does not need it). Without internet access but on your home Wi-Fi, your iPhone still reaches your computer if, in Devices, you chose “To devices on the local network” instead of “On this computer only”, the default; and Twoody still controls your Home Assistant.

Downloading a model, the first one included, or installing MLX; getting the app's updates; using online connectors, and refreshing Outlook calendar or contacts; signing in to the Twoody account, finding your conversations on your other devices, reaching your computer from the iPhone away from home, or sharing a folder. Web search too: if it is on while you are offline, the message does not go; turn it off until you are back online.

A large language model (LLM) is the kind of AI behind ChatGPT. A local LLM runs on your own computer instead of a company's servers: your questions are processed on your machine, and it works even offline. Twoody installs one for you, in one click.

8 GB is enough: Twoody then suggests Qwen3.5 4B. It suggests Qwen3.5 9B, which gives finer answers, from 16 GB on a Mac with Apple silicon, and from 32 GB on an Intel Mac, on Windows and on Linux (on Windows, from 16 GB with a graphics card that has 10 GB of its own). From 24 GB, it also offers the larger Qwen3.6 35B-A3B. In Models, each of the 57 models of the catalogue says the memory it needs, and the “Fits in this computer’s memory” filter shows only those that fit.

Yes. Twoody runs on macOS 13.3 (Ventura) or later, on Apple Silicon and on Intel. On an Intel Mac the model runs on the processor: it works, more slowly.

The same app runs on Mac (macOS 13.3 or later), Windows (10 and 11, 64-bit) and Linux (x86_64). The Mac remains the most complete version: only it reads scanned PDFs and the Mac's Calendar and Contacts, dictates with macOS speech recognition and, on Apple silicon, can use MLX. On iPhone and Android, the app uses the model of your computer. All of it is in private beta, not yet a public download: leave your email to join the waitlist.

On your other computers, Mac, Windows or Linux, yes, with the optional Twoody account, which is free: the conversations you keep on it open on the computers signed in to the same account, and private conversations stay on their computer. On a phone, the iPhone and Android apps, in private beta, work with Twoody on your computer and let you pick up your conversations.

Yes, with the optional Twoody account: a computer running Twoody without a screen, such as a Raspberry Pi 5, joins your account with a code, and “Automatic” sends each question to the one that will answer fastest. A model is never split across machines: each question goes to a single computer. By default, another computer receives only the conversation, without your memories or your documents, and what your computers exchange is encrypted end to end. The desktop app can welcome such a computer, but installing Twoody on a headless server is not published yet.

They are read on your computer, by a model installed on it, not by an online AI service. In a conversation you keep on your optional Twoody account, your messages, the answers and the passages Twoody read are kept on the account, encrypted end to end on your devices, so that you find them on your other devices: Osmove holds no key that can read them. A private conversation is kept on the computer only. Web search, off by default, sends your search to the engine you chose; a connector you add receives what you allow it to receive.

By default, Twoody suggests Qwen3.5, published by Alibaba's Qwen team under the Apache 2.0 licence, which we convert for llama.cpp, the engine inside Twoody. The other models of the catalogue are those of their publishers (Google, Mistral AI, Meta, OpenAI, DeepSeek…), each under its own licence, shown in the app. Twoody downloads them from Hugging Face and checks each file's fingerprint before using it. Once installed, a model is a file on your computer, and the engine only listens to your computer.

Twoody's interface, on the computer and on the iPhone, is in ten languages: English, French, German, Spanish, Italian, Portuguese (Brazil), Dutch, Japanese, Chinese (Simplified) and Korean; on the computer, choose yours in Settings › General › Language. Twoody answers in the language you write in, and the model understands many more: Qwen3.5 supports 201 languages and dialects.

Yes: Twoody is free for personal and professional use, and your clients' documents are read on your computer, by the model installed on it. For a client's file, create a walled-off project: its documents, conversations and notes serve only in that project. Or open private conversations for it: they are kept on that computer only.

A language model can be wrong, even when it cites its sources, and a model that runs on a personal computer is smaller than the large online models. Check its answers, especially before a professional decision.

Twoody keeps them on your computer, in its data folder, in an encrypted database whose key stays in the system's keychain (on Linux, when a keyring is available). It makes no backup of its own: a backup of your computer includes that folder, and to open it again on another computer, that keychain must be restored there too. The conversations you keep on your optional account are there too, encrypted end to end, and open on your other devices.

No. The app is free, but its code is not published. Models keep their own licence (Apache 2.0 for Qwen3.5) and the llama.cpp engine is under the MIT licence: Settings › About names them, and their texts ship inside the app.

Follow the two steps on this page: the app, then its data folder.

In the app, Settings › About › “Write to us”, or at support@twoody.com. Please do not send your clients' documents by email.