The gallery holds complete agents, ready to apply and adjust. They all go through the same validation as publishing — a template that would not publish does not make it into the gallery.
How each template presents itself
| Information | What it says |
|---|
| What it does | One sentence |
| Use case | The real situation it serves |
| Configuration time | An estimate after applying it |
| What it returns | The result's fields |
| What it requires | Tools, skills, addresses, or another agent |
| Availability | Ready to use · Needs configuration · Unavailable · Not confirmed |
A template only appears as ready to use when everything it requires exists and is allowed in your space. When something is missing, it says what. When it cannot be confirmed, it says that — and does not become "ready".
The ten templates
Text
| Template | What it does | Configuration |
|---|
| Summarize a text | Takes a text and returns a summary of the main points | ~1 min |
| Extract information | Reads free-form content and returns the fields you declare | ~3 min |
| Create content with approval | Generates a text and only returns it once someone approves | ~4 min |
Triage
| Template | What it does | Configuration |
|---|
| Classify a request | Reads a request and says which area it belongs to and at what priority | ~2 min |
| Analyze customer feedback | Returns sentiment, theme, urgency, and what to do | ~2 min |
Integration
| Template | What it does | Configuration |
|---|
| Call an API | Makes an HTTP call to a system and returns the response | ~5 min |
| Research a topic | Looks up current information and returns a summary with sources | ~2 min |
Periodic
| Template | What it does | Configuration |
|---|
| Periodic report | Prepares and analyzes a period's data and returns the report | ~5 min |
Orchestration
| Template | What it does | Configuration |
|---|
| Route a request | Classifies and forwards the urgent ones to another agent | ~6 min |
| Orchestrate another agent | Prepares the data, calls another published agent, and uses the result | ~4 min |
What you have to choose after applying
Some templates leave choices open — and Studio marks each one:
| Pending choice | In which templates |
|---|
| The API's address | Call an API, Periodic report |
| The agent to call | Route a request, Orchestrate another agent |
| The called agent's version | The same |
A template with a pending choice never appears as ready.
Common requirements
| Requirement | Why |
|---|
| The organization's service account | AI tasks run on the organization's behalf |
| Web search allowed | For the research template |
| An address allowed in the space | For the templates that call APIs |
| Another published agent | For the orchestration templates |
Recommended flow
- Apply the template closest to what you want.
- Adjust the objective, input, and result.
- Test.
- Activate in
development and run it a few times.
- Promote.
Next steps