Getting Started With Staffless
Set up your first Staffless workspace, connect tools, and run your first production-grade AI task in minutes.
Create Your First Workspace
A workspace is your isolated agent runtime. It has its own chat history, skills, external connections, scheduled tasks, and usage tracking.
Use a clear workspace name tied to one goal, for example 'daily-content' or 'client-success'. You can always create additional workspaces for separate workflows.
- Open Dashboard and create a new workspace.
- Wait until provisioning status is ready.
- Open Chat and send a simple test prompt.
Choose The Right Model
Start with MiniMax M2.7 for a strong price-to-performance baseline. Switch to higher-end models only for tasks that require deeper reasoning or stricter output quality.
Model choice affects speed, cost, and style. Keep your default model stable per workspace so behavior stays predictable.
Run Your First Real Task
Avoid vague tests like 'say hello'. Give one concrete operational task with a clear completion signal, then review output quality in the Output tab.
If the task will repeat, move it into a skill and schedule it as a cron. That turns one-off prompting into repeatable automation.
- Task prompt with explicit outcome and constraints.
- Save prompt logic into a skill.
- Create a scheduled task once the result is reliable.
FAQ
How many workspaces should I start with?
Start with one workspace per business function. This keeps context clean and makes cost attribution easier.
Can I change model later?
Yes. Model selection is flexible and can be changed per run or per scheduled task.