Deployments are how you can expose your workflows as APIs! Select your
version and machine. Then select either “production” or “staging” for
your deployment
You should see a deployment popup shortly in the bottom left, this means your
deployment is live!
Then you can intilize your client and create a run
Copy
Ask AI
// Initialize your clientconst client = new ComfyDeployClient({ apiToken: process.env.COMFY_DEPLOY_API_KEY!,});// Create a run via deployment idconst { run_id } = await client.run({deployment_id: "<YOUR DEPLOYMENT ID>", // these inputs will change based on your situatoin inputs: { "positive_prompt": "", "negative_prompt": "" }});// Check the status of the run, and retrieve the outputsconst run = await client.getRun(run_id);
Dependencies are the models, static assets and custom nodes that
you need to run your workflows. We figure this out based on your local
environment and collect them so that you can get up and running quickly.
Currently we don’t get all the dependencies but we do get a lot of the obvious
ones out of the way.
Public share is a way that you can quickly share your workflows with your
teammates or prospective clients. Selecting “public”, it will show up in your “deployments” section.
When you click the public share tab in the deployment section, a popup with
fields to customize your Share page will show up. You’ll be able to add a -
description of your workflow, - a “cover photo” to showcase. - The title of
the page will be the name of the workflow. - Users will be able to test out
the workflow, with the inputs you defined
Your view share page will look something like thisCheck out this examples page for Comfy Deploy workflows! (They use sharepages!)
External inputs are how you define the inputs of your deployed APIs! This is
done by connecting Comfy Deploy’s “External nodes” to your existing workflows.
You can find them by searching “external” for in your ComfyUI
Workflow observability means being able to undersatnd what happens on your
runs. What you see here is the inputs given, a link to the logs as well as the
final output.
Also you are able to see the time spent in the different operations. As well
as the fine grained breakdown if you hover over the Duration.