AI workflows you can build, run and understand
Hot wasl brings models, data and actions into one visual workspace. Connect the steps, run your workflow and inspect what happened at every node.
See how the pieces connect
An illustrative workflow: read a page, process its content, then deliver the result. Scheduled runs require an external scheduler.
- node types
- 28
- models in the catalog
- 25
- providers in the catalog
- 11
- interface languages
- 2
From input to action, on one canvas
Build visually
Drag nodes onto the canvas and connect them. Start with a template or draft a graph from an Arabic or English description, then review and edit every step.
Use the right model
Choose a model on each AI node and connect an OpenAI-compatible provider or gateway. Actual model access depends on the configured endpoint and its credentials.
Branch and process lists
Route values through conditions and categories. Process list items individually and combine the results without hiding the logic in a separate workflow.
Inspect every run
Read each node's inputs, outputs, logs, timing and credit usage. Follow the trace to understand a result or locate a failed step.
Connect your tools
Read web pages, call HTTP endpoints and send results to Slack or a webhook. Trigger workflows from your own application through the REST API.
Control model credentials
Attach your own model key to an AI node. Those model calls use zero platform credits; the provider bills you separately and tools such as web-page reads still consume credits.
A practical path to your first workflow
- 1
Choose a starting point
Open a template or create an empty workflow. Decide what input you will provide and what output you need.
- 2
Configure and test
Set the prompts, inputs and credentials. Run with sample data and review the trace before relying on the result.
- 3
Connect it to your work
Use the workflow manually or call it from your app. Publish webhook workflows before sending requests to their URL.