> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ringg.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> To search or read these docs, connect the Ringg docs MCP server at https://docs.ringg.ai/mcp (no key needed; tools search_ringg_ai and query_docs_filesystem_ringg_ai) instead of browsing pages one by one. Setup for each client: https://docs.ringg.ai/resources/mcp-server/docs-mcp-server/mcp-for-ai-assistants.md
> The dashboard labels agents "Assistants"; these docs say "agent". Both mean the same object, identified by agent_id.
> API base URL: https://prod-api.ringg.ai/ca/api/v0. Send the workspace API key in the X-API-KEY header, from server-side code only.
> Placing a call dials a real phone number and spends credits. Confirm the number with the user before calling.
> To act on a Ringg workspace (place calls, edit agents, read transcripts), use the separate Ringg workspace MCP server at https://mcp.ringg.ai/mcp: https://docs.ringg.ai/resources/mcp-server/workspace-mcp-server/overview.md

# Test and convert multi-prompt agents

> Test a Ringg AI multi-prompt agent, convert a single-prompt agent into a flow, and build or edit flows from an AI assistant through the MCP server.

## Test a multi-prompt agent

<Steps>
  <Step title="Publish or discard pending changes">
    While the agent has unpublished changes, the top bar shows **Discard** and **Publish** instead of **Test assistant**. Publish them first. See [Publish changes on a multi-prompt agent](/documentation/build-agents/test-and-improve/drafts-and-versions#publish-changes-on-a-multi-prompt-agent).
  </Step>

  <Step title="Choose a test mode">
    Click **Test assistant** and pick **Phone call**, **Web call** or **Live chat** ([test options](/documentation/build-agents/test-and-improve/test-an-agent#test-options)). Phone call opens a dialog; web call and live chat run in the right panel so the canvas stays visible.
  </Step>

  <Step title="Watch the flow">
    During the test, the canvas highlights the node the conversation is on and marks the path taken so far. Extracted values appear on the nodes as they are captured.
  </Step>

  <Step title="Review the run">
    When the call ends, the path stays on the canvas with a **Test call ended** pill. Click **View summary** to see the **Path** and **Extracted variables**.

    <Check>The path on the canvas matches the branch you expected, and the extracted values are correct. The call is also in [Logs](/documentation/monitor/logs).</Check>
  </Step>
</Steps>

A test runs the version shown in the editor. For test modes and caller variables, see [Test an agent](/documentation/build-agents/test-and-improve/test-an-agent).

## Convert a single-prompt agent

Where conversion is available, a single-prompt agent shows **Convert to multi-prompt** under **More** in its editor sidebar; its tooltip shows the prompt's token count. It isn't offered to every workspace: if you don't see it, contact Ringg.

<Steps>
  <Step title="Open the dialog">
    Click **Convert to multi-prompt**. The dialog **Convert to multi-node agent** explains that your current agent will remain unchanged.
  </Step>

  <Step title="Start the conversion">
    Click **Upgrade**. Progress streams in: plan generated, start and end nodes created, then each Agent, Speak and Router node and its connections. You can close the dialog; the conversion keeps running.
  </Step>

  <Step title="Open the result">
    When the dialog title changes to **Multi-node agent is ready**, click **Open new agent**. The conversion creates a separate multi-prompt agent. Review its nodes, then test it before you route real calls to it.

    <Check>The new multi-prompt agent opens on the canvas. Your original single-prompt agent is unchanged.</Check>
  </Step>
</Steps>

## Build flows with AI tools

The Ringg [MCP server](/resources/mcp-server/workspace-mcp-server/overview) can create multi-prompt agents and edit their nodes and connections as a draft, then publish them, so an AI assistant that supports MCP can build or change a flow for you. See [connect the MCP server](/resources/mcp-server/workspace-mcp-server/connect-an-assistant) and its [tools](/resources/mcp-server/workspace-mcp-server/tool-reference).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.