chat
<p>[TOC]</p>
<h5>简要描述</h5>
<ul>
<li>智能体chat接口</li>
</ul>
<h5>请求URL</h5>
<ul>
<li><code>https://openapi.tuniu.cn/agent/api/chat</code></li>
</ul>
<h5>请求方式</h5>
<ul>
<li>POST</li>
<li>application/json</li>
<li>流式输出接口</li>
</ul>
<h5>请求头</h5>
<table>
<thead>
<tr>
<th>参数名</th>
<th>参数描述</th>
<th>类型</th>
<th>是否必填</th>
</tr>
</thead>
<tbody>
<tr>
<td>apikey</td>
<td>开放平台访问秘钥</td>
<td>String</td>
<td>是</td>
</tr>
</tbody>
</table>
<h5>请求体</h5>
<table>
<thead>
<tr>
<th>参数名</th>
<th>参数描述</th>
<th>类型</th>
<th>是否必填</th>
</tr>
</thead>
<tbody>
<tr>
<td>sessionId</td>
<td>会话id信息（如果不填会自动生成新的，在响应里带回去）</td>
<td>String</td>
<td>否</td>
</tr>
<tr>
<td>requestId</td>
<td>标记单次请求的ID（如果不填会自动生成新的，在响应里带回去）</td>
<td>String</td>
<td>否</td>
</tr>
<tr>
<td>agentId</td>
<td>Agent的ID 机票：FlightTicketAgent 酒店：HotelAgent火车票：TrainTicketAgent</td>
<td>String</td>
<td>是</td>
</tr>
<tr>
<td>prompt</td>
<td>发送的消息</td>
<td>String</td>
<td>是</td>
</tr>
<tr>
<td>messages</td>
<td>用户自定义历史记忆（若同时传入sessionId和messages，则大模型优先使用messages中的内容，sessionId将被忽略，重新生成一个新的sessionId返回）</td>
<td>Message</td>
<td>否</td>
</tr>
</tbody>
</table>
<h5>请求示例</h5>
<pre><code class="language-json">{
&quot;sessionId&quot;: &quot;xxx&quot;,
&quot;requestId&quot;: &quot;xxx&quot;,
&quot;agentId&quot;: &quot;TrainTicketAgent&quot;,
&quot;prompt&quot;: &quot;明天从南京去厦门的火车票&quot;,
&quot;messages&quot;: [
{
&quot;role&quot;: &quot;user&quot;,
&quot;content&quot;: &quot;你好&quot;
},
{
&quot;role&quot;: &quot;assistant&quot;,
&quot;content&quot;: &quot;请问有什么需求&quot;
}
]
}</code></pre>
<h5>响应体</h5>
<table>
<thead>
<tr>
<th>参数名</th>
<th>参数描述</th>
<th>类型</th>
<th>是否必填</th>
</tr>
</thead>
<tbody>
<tr>
<td>statusCode</td>
<td>响应码（200成功）</td>
<td>Integer</td>
<td></td>
</tr>
<tr>
<td>message</td>
<td>响应描述</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>sessionId</td>
<td>会话id信息</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>requestId</td>
<td>请求id信息</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>info</td>
<td>具体的响应内容</td>
<td>List<Response></td>
<td></td>
</tr>
<tr>
<td>apikey</td>
<td>开放平台访问秘钥</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>usage</td>
<td>模型信息（只在流式输出最后一段内容里携带该信息）</td>
<td>Usage</td>
<td></td>
</tr>
</tbody>
</table>
<p>Response</p>
<table>
<thead>
<tr>
<th>参数名</th>
<th>参数描述</th>
<th>类型</th>
<th>是否必填</th>
</tr>
</thead>
<tbody>
<tr>
<td>type</td>
<td>文本类型</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>content</td>
<td>具体响应内容</td>
<td>String</td>
<td></td>
</tr>
</tbody>
</table>
<p>Usage</p>
<table>
<thead>
<tr>
<th>参数名</th>
<th>参数描述</th>
<th>类型</th>
<th>是否必填</th>
</tr>
</thead>
<tbody>
<tr>
<td>model</td>
<td>模型信息</td>
<td>String</td>
<td></td>
</tr>
<tr>
<td>inputTokens</td>
<td>输入token数量</td>
<td>Integer</td>
<td></td>
</tr>
<tr>
<td>outputTokens</td>
<td>输出token数量</td>
<td>Integer</td>
<td></td>
</tr>
</tbody>
</table>
<pre><code class="language-json">{
&quot;statusCode&quot;: 200,
&quot;message&quot;: null,
&quot;sessionId&quot;: &quot;20250343222&quot;,
&quot;requestId&quot;: &quot;34983848394&quot;,
&quot;info&quot;: [
{
&quot;type&quot;: &quot;analysis&quot;,
&quot;content&quot;: &quot;- 正在调用 `火车票小助手` 为您服务\n&quot;
}
],
&quot;usage&quot;: {
&quot;model&quot;:&quot;qwen2.5-32b-instruct&quot;,
&quot;inputTokens&quot;:1800,
&quot;outputTokens&quot;:200
}
}</code></pre>