NOW IN OPEN BETA

Anaconda Desktop

Get environments, models, and endpoints in one place, with security and governance built in.

Read Documentation
Windows64-bit installer macOS64-bit installer Linux (x86)Ubuntu/Debian 64-bit installer Linux (ARM64)Ubuntu/Debian 64-bit installer

Built for How You Actually Work

AI models, Python environments, apps, and endpoints. One local interface.

Illustration of the Anaconda Desktop AI model browser showing the NVIDIA-Nemotron-Nano-9B-v2-Base model card with quantization, size, and RAM details, alongside a system resource panel showing CPU and RAM usage, and a chat input — illustrating model discovery, benchmarking, and local inference.

Find the right AI model for the job

Browse enterprise-ready AI models, pre-vetted for security and licensing. Filter by task, hardware compatibility, quantization, and RAM requirements to find models that actually run on your machine, not ones you have to troubleshoot. Compare benchmarks side-by-side and start chatting or serving in minutes.

Build and run locally

Spin up API endpoints from any downloaded model with one click. Connect to your applications, monitor CPU and RAM utilization, and view server logs—all without cloud API costs, and without data leaving your machine. Prototype and iterate locally, then deploy when you’re ready.

Illustration showing an API inference server running locally at 127.0.0.1:8081 alongside a connected Jupyter Notebooks instance, demonstrating Anaconda Desktop's ability to spin up and manage local model endpoints without cloud dependencies.
Illustration of Anaconda Desktop's environment creation options, showing a floating menu with choices to Import, Duplicate Existing Environment, Create Manually, or use Quick Start Environments, connected by a line to a dark card displaying the anaconda-ml-ai environment with 24 packages and a size of 6.44 GB.

Keep projects clean and reproducible

Create, clone, and manage conda environments from a GUI instead of the command line. Start building immediately with ready-to-use environments or build your own from scratch. Search and install packages without leaving the interface. And with import and export configs, you can keep projects reproducible across machines—no manual setup required.

Launch your favorite tools

Open Jupyterlab, Spyder, and other applications directly from Desktop. Install, update, and manage your applications alongside your environments and AI models, all from one interface.

Illustration showing Anaconda Desktop as a central hub for launching data science tools, with Jupyter Notebook, JupyterLab, and Spyder application icons arranged around a download icon, representing one-click installation and launch of favorite development environments.

Anaconda Desktop Frequently Asked Questions

What is Anaconda Desktop?

Anaconda Desktop is a free desktop application for local AI development. It provides a curated AI model catalog, local model inference, environment and package management, and application launching, all in a single interface.

Anaconda Desktop retains Navigator’s environment management and application launching capabilities and adds AI model discovery and local model servers.

Anaconda Desktop includes a curated model catalog with models vetted for security and licensing. You can browse and filter models by task, hardware compatibility, quantization format, and RAM requirements.

Yes. Desktop lets you compare models side-by-side across benchmarks for reasoning, math, code, and function calling. You can filter by hardware requirements, license type, and task to find models that match your use case and machine.

Yes. All inference runs on your machine. Your prompts, outputs, and data never leave your device.

Yes. Anaconda Desktop provides a graphical interface for creating, cloning, and managing conda environments.

You can launch JupyterLab, Spyder, and other supported applications directly from Desktop. Applications are installed and managed alongside your environments and AI models.