> For the complete documentation index, see [llms.txt](https://docs.zerve.ai/zerve-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.zerve.ai/zerve-docs/superagents/create-a-superagent.md).

# Create a superagent

Create a superagent from a mission, review its plan and approve it to start work.

You create a superagent by describing its mission. The superagent drafts a plan, you review and edit it, and it starts work when you approve.

## Before you start

The planner looks at what already exists in your organization when it drafts a plan: your [data connections](/zerve-docs/data-and-assets/connections.md), [constants and secrets](/zerve-docs/data-and-assets/constants-and-secrets.md) and [skills](/zerve-docs/agents-in-the-workspace/skills.md). Setting these up first gives you a more grounded plan. If something's missing, the superagent will ask for it as a [decision](/zerve-docs/superagents/decisions-and-recommendations.md).

## Describe the mission

1. Open the product switcher and choose **Superagents**.
2. On the home page, describe the role or mission you want the superagent to own.
3. Optionally, attach files or use voice dictation (available in desktop browsers).
4. Select **Start**.

You can also select **New Superagent** from the **Your superagents** list.

### Writing a good mission

The home page suggests writing it "the way you would say it out loud." A few things help:

| Do                                         | For example                                                                                                 |
| ------------------------------------------ | ----------------------------------------------------------------------------------------------------------- |
| **Describe the outcome, not the steps**    | "Find which momentum signals still work in European equities, and why" rather than a list of queries to run |
| **Say what an ongoing version looks like** | "…and every Monday, tell me what changed"                                                                   |
| **Name the data you expect it to use**     | "…using our positions data in Snowflake and public market data"                                             |
| **Say what you'll judge it on**            | "…a signal only counts if it holds up out of sample"                                                        |

To see complete examples, select one of the **Try this:** options on the home page.

## How the plan is drafted

A planner reads your mission and your organization's resources, and can search the web. It then drafts:

* A **name** for the superagent
* One or more **objectives**, each with a **KPI** and an **autonomy level**
* Suggested **schedules**, such as a weekly briefing
* Any **recommendations**, meaning decisions or inputs it needs from you

The planner is built to avoid overreach. It usually proposes a single objective, at most two or three, and leads with the cheapest step that tests whether the mission is achievable.

While it plans, the screen shows short status lines about your mission. If planning fails, you'll see an error and can delete the draft and try again.

## Review the plan

When the plan is ready, the screen shows **Here is what I plan to do**. Your mission is shown for reference and can't be changed.

You can:

| Change                  | How                                                                           |
| ----------------------- | ----------------------------------------------------------------------------- |
| **Edit an objective**   | Change its text directly                                                      |
| **Add an objective**    | Select **Add a step**                                                         |
| **Remove an objective** | Remove it from the list                                                       |
| **Set autonomy**        | Choose **Ask me first**, **Balanced** or **Full autonomy** for each objective |
| **Edit KPIs**           | Change a KPI's name, target or direction                                      |
| **Remove a schedule**   | Remove any proposed schedule you don't want                                   |

Proposed schedules stay paused until you approve the plan.

For more on objectives, KPIs and autonomy, see [Objectives, KPIs and autonomy](/zerve-docs/superagents/objectives-kpis-and-autonomy.md).

## Approve and start

Select **Approve & start**. Three things happen:

1. The superagent becomes **active**.
2. Its schedules start.
3. A new conversation opens, where the superagent tells you what it's starting on first, then starts it.

From here, the superagent works continuously. See [Monitor a superagent](/zerve-docs/superagents/monitor-a-superagent.md).

## Next steps

After you create a superagent, you can adjust it in its [settings](/zerve-docs/superagents/settings.md): its name, purpose and colour, persona and guardrails, models, channels and daily budget.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.zerve.ai/zerve-docs/superagents/create-a-superagent.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.