We are thrilled to announce a major update to RunComfy, Workflow "Cloud Save" and "Sharing".
1. Why do we need this update?
This update solves the following problems,
- As a ComfyUI user, you may encounter challenges when trying to save a workflow along with its environment for future reuse. The current system lacks the ability to save the complete setup, making it challenging to reproduce results later.
- As a ComfyUI user, you may find that the environment of each workflow is not entirely independent. This lack of isolation can cause conflicting dependencies between workflows, requiring manual resolution each time.
- As a ComfyUI user, if you attempt to reproduce a workflow shared by others, you might face red errors indicating missing nodes, models, or libraries, as the necessary components aren't included or configured properly.
- As a ComfyUI workflow creator, you may need to provide additional support and guidance to users who are trying to set up and run your shared workflows.
2. How does this update solve the problems above?
Based on your valuable feedback, we are introducing the following two powerful new features to tackle these problems once and for all:
2.1 Workflow "Cloud Save"
With Cloud Save, you can now seamlessly save your workflow along with the entire GPU machine environment to the cloud. This includes any changes while working with your workflow:
- OS update (drivers and packages)
- Python libraries
- ComfyUI releases
- Custom Nodes
- New Models
All your configurations will be saved alongside any workflow you are working on, so you can pick up right where you left off—without losing any progress. You can also reproduce the same results in the future, as the entire setup is preserved.
Furthermore, Cloud Save allows you to save different workflows with their respective environments, eliminating the need to manually resolve conflicts between workflows.
Here is how you use it:
- After “Run Workflow”, you can use the "Cloud Save" feature to save this workflow alongside the entire environment.
- After you have successfully saved your work and stop the machine, you will find your saved workflow in "My workflows." Just run it again and you will find the workflow and environment is exactly the same when you saved it.
2.2 Workflow "Sharing (Beta)"
Collaboration has never been easier! With our new Sharing feature, you can now share your saved workflows with anyone via a simple link.
- No setup required: The person receiving your link can open and run the workflow immediately without any additional configuration (They still need to provide their own input files, such as image/video, but the prompts/settings/nodes/models are all saved in the workflow).
- Privacy assured: Your inputs, outputs, and private model folders remain secure and are not included in the shared workflow. Only the Cloud Environment, ComfyUI, and any necessary models referenced in your workflow will be shared.
Here is how you share the workflow,
3. What about existing users’ data?
We are committed to make the migration from old infra to new as smooth as possible. For all existing users on Pro Plan and active users on Hobby Plan, you will find a “Legacy” Workflow that contains all of your previous custom_nodes. You should be able to run that “Legacy” workflow and reproduce the previous working environment in RunComfy.
What to do if the “Legacy” is not working?
- Please contact hi@runcomfy.com or DM @henlee in discord, we will look into the migration issues as soon as possible.
- Your models/input/output will stay the same when you run different workflows, thus a much faster way to get back on track is to run a default workflow and install whatever nodes you need and save that workflow as yours.
4. Key Concepts update in RunComfy
Terms | Explanation |
Assets | Your private drive contains your own models, input, and output. |
RunComfy ComfyUI Workflow | A reproducible ComfyUI workflow contains the workflow json, the running OS/python environment, the ComfyUI, the necessary custom nodes, and the necessary models. |