Ollama
Ollama is an open-source tool for running large language models (AI chat models) on your own computer without an online service. Models are downloaded with a single command and the chat runs directly on your machine.
github.com · Official website
- ✓ Official source (winget)
- ✓ SHA256 taken from the official repository
- ✓ Last checked: Nov. 4, 2026
- This is not a virus scan. Check on VirusTotal
- Version
- 0.40.14
- Release date
- Nov. 3, 2026
- License
- MIT
- Installer type
- inno
How do I use this?
Type Terminal in the Start menu and open it, paste the command (Ctrl+V or right-click) and press Enter. It uses winget, Windows' own package manager, to install the program from its official source.
winget install --id Ollama.Ollama -eSHA256 (file verification)
e069c7cc7d4808f77684e81efa0b1f853a5190b4b8f070bc8fd63041180e3ebagithub.com · Official website
- ✓ Official source (Homebrew)
- ✓ SHA256 taken from the official repository
- ✓ Last checked: Nov. 4, 2026
- This is not a virus scan. Check on VirusTotal
- Version
- 0.40.14
- License
- MIT
- Installer type
- zip
How do I use this?
This command needs Homebrew; if you don't have it, first run the install command from brew.sh. Then open Terminal (Applications > Utilities > Terminal), paste the command and press Enter.
brew install --cask ollama-appSHA256 (file verification)
9047386fb6811f54fc98b12ce4fea42a6ee5ce9e1bbc0c8196499f347054bcb5We couldn't find a Linux version of Ollama in the official repositories; check the publisher's website.
You can try these instead:
We couldn't find a Android version of Ollama in the official repositories; check the publisher's website.
Ollama 0.40.14: what's new?
Highlights
- Download and run open models such as Gemma, Llama, Mistral and Qwen with one command
- Chats stay on your computer
- Local API that other programs can connect to
- Uses your graphics card when available for extra speed
- Commands to list, remove and update models
Who is it for?
For enthusiasts and developers who want to run AI models on their own computer for privacy or experimentation.
Good to know before installing
- Basic use is from the command line; if you prefer a graphical app, LM Studio may suit you.
- Large models need strong hardware and run slowly on weak PCs.
How to install
Windows
- Click the official download button on this page or use the winget command.
- Run the installer.
- Open Terminal and type ollama run followed by a model name to download your first model.
- Once the model is downloaded, the chat starts in the same window.
macOS
- Use the Homebrew command or download from the official site.
- Move the app to the Applications folder and open it.
- In Terminal, type ollama run and a model name to download your first model.
Alternatives
- LM Studio
- GPT4All
- Jan
Frequently asked questions
Is Ollama free?
Yes. It is open source (MIT license) and free. Each model has its own license.
What kind of computer do I need?
It depends on the model size. Small models run on an ordinary PC; large models need lots of memory and a strong graphics card. Models can take up several gigabytes.
Does it need an internet connection?
Only while downloading a model. After that, the chat runs on your computer.