Feature
A visual AI workflow builder
Build the workflow your crew will follow. Start from a description, then adjust triggers, steps, and tools on the canvas before anything runs in production.
The problem
A chat box can suggest the next email. It does not remember the path, the tool, or the approval. Teams end up re-explaining the same process. A workflow builder keeps that process as a version you can test and deploy.
A concrete example
You tell the Architect: when a new Gmail message matches a search, read it, draft a reply, and wait before sending. The draft crew shows those steps. You can change the search, the instructions, or the tool on the canvas. Deploy pins that version. Later edits create another version instead of silently changing production.
How it works, and the limit
The Architect and AI suggestions spend AI credits. The hold is calculated before the model is called. If the workspace cannot cover that hold while billing is enforced, the request is refused and nothing is sent to the provider.
A normal crew run after deploy uses the model key the workspace saved. That execution is not metered as CrewX24 credits.
Questions
Can I edit the workflow after the Architect drafts it?
Yes, on a paid plan. The canvas and the conversation edit the same crew definition. The free trial cannot open a blank canvas or edit the pre-built crew.
Does deploying overwrite history?
Deploy puts one version in production. Runs that already started finish on the version they began with. You can roll back to a version that was in production before.