ML,
without the FML

Open source platform for machine learning dev environments. SSH into GPU machines across cloud providers.

Join Discord
Unweave Screenshot
dcf logospeedinvest logopioneer logo

Get started in seconds

Get up and running with an ML environment on an instance of your choice

1. Install

$ brew install unweave/unweave/unweave

2. Initialize

$ unweave link <project-id>

3. Connect

$ unweave ssh --provider unweave --gpu-type a100_x4

Choose your provider:

Choose your runtime:

Transparent Cloud Costs*

Unweave converts cloud provider spaghetti pricing into something parseable by mere mortals

Unweave

H100 PCIe
$4.30/hr
A100 40GB
$2.10/hr
RTX A6000
$1.30/hr
A40
$1.30/hr

LamdaLabs

A100 40GB
$1.10/hr
RTX A6000
$0.80/hr
A10
$0.60/hr
Quadro RTX 6000
$0.50/hr

More coming soon ...

A100 80GB
$2.70/hr
A40 48GB
$1.90/hr
RTX 6000 32GB
$1.50/hr
A10
$0.90/hr
A100 80GB
$2.70/hr
A40 48GB
$1.90/hr
RTX 6000 32GB
$1.50/hr
A10
$0.90/hr
A100 80GB
$2.70/hr
A40 48GB
$1.90/hr
RTX 6000 32GB
$1.50/hr
A10
$0.90/hr

*We update the exact pricing per provider at runtime. We do not charge a mark up on pricing of cloud providers. Unweave is free to get started by bringing your own cloud provider.
See here for the platform pricing.

Defaults that just work

Unweave chooses sane defaults that just work, while letting you tweak to your heart's content.

Run from your terminal

A cross platform CLI that fits your workflow whether locally or in a browser IDE. It works just as well in CI/CD pipelines.

Manage from the Browser

The Unweave Dashboard gives you a convenient realtime view of your projects and sessions.

Zero setup remote GPUs

Who want's to configure cloud infra when you could doing `model.fit()`. Unweave sets up your cloud dev environment for you.

Work across cloud providers

GPUs are in short supply. What if you could search across cloud providers for the best price and availability?

Sync code and data with Git

(Coming soon) Automatically version your code, data and artifacts with Git, with inbuilt CI/CD integrations.

Clone code and data seamlessly

(Coming soon) Start ML experiments directly from your Git repo. Unweave will automatically clone your code and data.

Built for flexibility

Develop with tools you're already used to. Unweave is built on top of plain old SSH and Git.

Unweave Screenshot

Your code editor

Write code in the IDE you're already used to. Whether on your local laptop or in a browser IDE. Unweave

Unweave Screenshot

Unweave CLI

Setup dev environment that integrate with your IDE straight from your terminal with the Unweave CLI.

Unweave Screenshot

Unweave Dashboard

If you prefer a more visual interface, everything you can do with the CLI the Unweave Dashboard supports too.

Join the community

Follow the development on Github and join the community to help shape the product.