标签增强
- 把简短的风格标签扩写成更丰富的组合，异步任务
- 结果在
data.result.upsampled_tags - 扩写后的标签可直接喂给「生成音乐」
认证
All endpoints authenticate with a Bearer token.
Get an API key: open the console and create one.
Add it to the request header:
Authorization: Bearer YOUR_OAIRoute_KEY请求参数
音频模型，当前只支持 suno，不传时也默认 suno。
待增强的风格标签，例如 pop。
缺失时提交阶段直接返回 400，且不扣费。
本端点没有版本维度。传 version 会被丢弃，且不影响计费。
The audio model, currently only suno; defaults to suno when omitted.
The style tag to enhance, for example pop.
Missing returns a 400 straight away at submission, with no charge.
This endpoint has no version dimension. Passing version is discarded and doesn't affect billing.
响应
响应状态码，成功为 200。
任务状态，提交后为 submitted。
任务唯一标识，用于轮询。
The response status code; 200 on success.
The task status; submitted after submission.
The unique task identifier, for polling.
取结果
用 GET /v1/music/tasks/{task_id} 查询，增强后的标签在 data.result.upsampled_tags。
提交成功只代表任务已被受理。增强结果必须通过查询接口取回，不能指望提交响应里直接带。
Query with GET /v1/music/tasks/{task_id}; the enhanced tags are in data.result.upsampled_tags.
A successful submission only means the task was accepted. The enhanced result must be fetched through the query endpoint — don't expect it to ride along in the submission response.
为什么需要这一步
风格标签写得太短，模型能发挥的空间反而小。比如只写 pop，生成结果可能偏任何一种流行乐。扩写成 pop, female vocal, bright piano, 120 bpm, radio-ready 之后，方向就明确得多。
典型用法是把它当成「先把想法写具体」的助手：
- 你有模糊的方向，写一个词
- 调本接口扩写
- 把扩写结果粘进「生成音乐」的
tags或style
Style tags written too briefly leave the model little room. Write only pop, for instance, and the result might lean toward any kind of pop. Expand it to pop, female vocal, bright piano, 120 bpm, radio-ready and the direction is far clearer.
The typical use is as an assistant for "make the thought concrete first":
- You have a vague direction and write one word
- Call this endpoint to expand it
- Paste the expanded result into "Generate music"'s
tagsorstyle
什么时候不用它
如果你已经能写出三个以上的具体标签，直接生成就行，不用多这一步。这个接口的价值在于「把模糊变具体」，不是「把具体变更多」。
| 情况 | 建议 |
|---|---|
| 只想到一个词 | 先扩写 |
| 已能写 3 到 5 个标签 | 直接用 |
| 想要特定风格 | 直接写，别让模型猜 |
扩写结果不一定完全符合你的意图，加了太多你没想到的元素是常见情况。拿到结果后先看一眼，把不想要的标签删掉再用于生成。
If you can already write three or more specific tags, just generate — no need for the extra step. This endpoint's value is "making the vague concrete", not "making the concrete more numerous".
| Situation | Advice |
|---|---|
| You can only think of one word | Expand first |
| You can already write 3 to 5 tags | Use them directly |
| You want a specific style | Write it out; don't make the model guess |
The expansion doesn't always match your intent — adding elements you didn't think of is common. Look the result over before using it, and delete tags you don't want.
错误码
For error response formats, status codes, and retry guidance see Error codes and rate limits; for troubleshooting steps on common errors see the FAQ.
使用示例
提交一个任务并拿到 task_id：
提交成功只表示任务已受理。真正的产物要拿返回的 task_id 去任务查询接口轮询，status 变成 completed 之后才有结果。
Submit a task and get a task_id:
A successful submission only means the task was accepted. To get the actual output, poll the task query endpoint with the returned task_id; results appear only after status becomes completed.