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

# List test conversations

> List your most recent Playground tests with the task they ran on and their last messages, to check what the agent answered.

When you try an agent in Playground, Nexor creates a test lead for that conversation. Test leads never show up in the lead lists, so this endpoint is the way to find them again. Use it to check what the agent answered in a test that went wrong, before you change its instructions.

It returns your most recent tests first, with the task each one ran on and its last messages.

## Before you start

* MCP keys need `leads:read`. Full-access REST keys need nothing else.
* Only Playground tests are returned. Real leads are never included.
* It is read-only. Calling it does not send anything to anyone.
* Errors return a stable code in `error`, such as `invalid_limit`. Map the code to your own copy.

## Choose what to read

| Parameter | Detail |
| - | - |
| `workflow_id` | Only tests that ran on one task (`workflow` in the API). A task from another account returns an empty list, not an error. |
| `limit` | 1 to 25. Default 10. |

## What you get back

| Field | Detail |
| - | - |
| `lead_id` | The test lead. Use it to read what the agent did in each turn of that test. |
| `name` | The name used in the test, when there is one. |
| `workflow_id`, `workflow_name` | The task the test ran on. |
| `started_at`, `last_activity_at` | When the test started and its most recent activity, in UTC. |
| `last_messages` | Up to the last 6 messages, oldest first. `inbound` is what the tester wrote and `outbound` is what the agent answered. |

## Read what the agent did

With the Nexor MCP, pass `lead_id` to `list_lead_turns` and `get_lead_turn`. Each turn shows the agent's reply, the tools it called, and any tool error. That tells you whether the problem is in the instructions, in a tool, or in something the platform does not do yet.

## Limits

* At most 25 conversations per request.
* Message text is cut at 280 characters. API keys, tokens, and signatures are masked. Anything the tester typed is returned as written, so treat it as conversation data.


## API Specification

The full API specification for this endpoint is available in the [documentation index](https://docs.getnexor.ai/llms.txt).


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