Skip to main content
Almost everything behind Omni is configurable: the model, provider, prompt, skills, tools, workspace context, and the agents it can delegate to. Tuning these settings gives you a personal agent that behaves consistently across your team’s real workflows, instead of a generic assistant.

​
What You Can Customize

You can configure:
  • Model
  • Provider
  • System prompt
  • Skills
  • Connected tools
  • Workspace files and folders
  • Available agents
  • Delegation behavior
  • Memory and reusable context
Omni customization settings

​
Behavior

The system prompt defines how Omni should behave. Use it to set tone, decision rules, workflow boundaries, output formats, and source-specific instructions. This is useful when Omni needs to follow a repeatable process, such as creating a morning brief, checking assigned tickets, or generating a shareable report.

​
Capabilities

Skills and tools define what Omni can do. A skill gives Omni workflow instructions. A tool or connector gives Omni access to external systems such as Gmail, Slack, Notion, GitHub, Calendar, or a custom API. Together, they let Omni move from answering questions to completing real tasks.

​
Context

Workspace files and memory help Omni remember what matters. Use them for schemas, templates, database IDs, user preferences, team-specific instructions, and reusable project context.

​
Delegation

Omni can coordinate other agents. Instead of asking users to choose the right agent every time, Omni can route work to the agent best suited for the task.

​
Best Practice

Start simple. Give Omni the tools, skills, and context required for the workflow you want it to run. Add more only when there is a clear need. A focused Omni is easier to trust, easier to debug, and easier for teams to adopt.