Agents Catalog
Runnit ships a catalog of AI agents that produce creative and strategic work inside your projects. Agents are launched from the app and run asynchronously, returning a summary and any files they create.
For ad-hoc questions and actions, use the page-aware Chat Assistant from the global header. It can work with the current page’s context, search accessible assets, generate media, and submit product feedback after you confirm the draft.
Built-in agents
Section titled “Built-in agents”| Agent | What it does |
|---|---|
| Copy Assistant | Scripts, taglines, campaign concepts, and long-form copy that respect brand tone. |
| Strategy Assistant | Go-to-market plans, channel strategy, messaging frameworks, and measurement roadmaps. |
| Synthetic Audience Tester | Tests creative with synthetic personas for sentiment and resonance. |
How agents run
Section titled “How agents run”Agents run inside Runnit itself. Every agent receives the same context for its run: the task instructions, the brief, team, and project details, and the files and brand collections it is allowed to read. It works step by step with Runnit’s tools, saves its deliverables into the job’s output collection, and finishes with a summary that references the files it created.
Organisation administrators can also create custom agents with their own goal, system prompt, and knowledge base, and can customise a built-in agent by editing its prompt, model, and tool access. The model agents use by default is set on the AI Usage page.