API 文档

快速开始

从拿密钥到跑通第一张图，一共三步。

第一步：创建密钥

进入「账号管理 → API 平台」，点右上角「创建密钥」，给它起个能认出用途的名字。创建完成时密钥会自动复制到剪贴板；之后随时可以在列表里点复制按钮再取一次。

密钥等同于你的账号，能直接花掉你的算力。列表页随时可以再复制，所以不必到处抄存副本——副本越多，泄露面越大。

第二步：看看有哪些模型

把密钥放进环境变量，先查一次模型列表——返回里的 id 就是后面下单时 model 参数要传的值。

bash
curl https://open.pikpikgo.com/v1/models \
  -H "Authorization: Bearer $PIKPIK_API_KEY"

第三步：出一张图

bash
curl https://open.pikpikgo.com/v1/images/generations \
  -H "Authorization: Bearer $PIKPIK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "赛博朋克风格的雨夜街道，霓虹倒影",
    "size": "16:9",
    "quality": "2k",
    "n": 1
  }'

默认是同步的：接口会一直等到图出完才响应，data[0].url 就是图片地址，不需要再查任务。单张通常 30 秒以上，请把客户端超时放宽到 300 秒。

json
{
  "created": 1786000246,
  "data": [
    { "url": "https://cdn.pikpikgo.com/ai/xxxx.png", "revised_prompt": "赛博朋克风格的雨夜街道，霓虹倒影" }
  ],
  "id": "2608101909450810293847",
  "object": "image.generation",
  "model": "img-std-1",
  "status": "succeeded",
  "n": 1,
  "credits": 10,
  "tasks": [
    {
      "id": "2608101909450810293847",
      "status": "succeeded",
      "data": [{ "url": "https://cdn.pikpikgo.com/ai/xxxx.png", "revised_prompt": "赛博朋克风格的雨夜街道，霓虹倒影" }]
    }
  ]
}

第四步：不想等？改成异步

请求里多加一个 async: true，接口立刻返回，status 是 queued、data 还是空的——图还没出。记下 tasks[0].id。

bash
curl https://open.pikpikgo.com/v1/images/generations \
  -H "Authorization: Bearer $PIKPIK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "赛博朋克风格的雨夜街道，霓虹倒影",
    "size": "16:9",
    "quality": "2k",
    "n": 1,
    "async": true
  }'
json
{
  "id": "2608101909450810293847",
  "object": "image.generation",
  "created": 1786000185,
  "model": "img-std-1",
  "status": "queued",
  "n": 1,
  "credits": 10,
  "tasks": [
    { "id": "2608101909450810293847", "status": "queued", "data": [] }
  ],
  "data": []
}

再拿这个 id 轮询任务，status 变成 succeeded 时，data[0].url 就是图片地址。生视频只有这一种用法。

bash
curl https://open.pikpikgo.com/v1/tasks/2608101909450810293847 \
  -H "Authorization: Bearer $PIKPIK_API_KEY"
json
{
  "id": "2608101909450810293847",
  "object": "image.generation",
  "created": 1786000185,
  "model": "img-std-1",
  "status": "succeeded",
  "prompt": "赛博朋克风格的雨夜街道，霓虹倒影",
  "size": "16:9",
  "quality": "2k",
  "credits": 10,
  "error": "",
  "data": [
    { "url": "https://cdn.pikpikgo.com/ai/xxxx.png", "revised_prompt": "赛博朋克风格的雨夜街道，霓虹倒影" }
  ]
}

完整示例

同步这条路 OpenAI 官方 SDK 可以直接用，只换 baseURL：

javascript
import OpenAI from 'openai'

const client = new OpenAI({
  apiKey: process.env.PIKPIK_API_KEY,
  baseURL: 'https://open.pikpikgo.com/v1',
  timeout: 300 * 1000,
})

// 同步出图：一次往返直接拿地址，超时放宽到 300 秒
const res = await client.images.generate({
  model: 'img-std-1',
  prompt: '赛博朋克风格的雨夜街道，霓虹倒影',
  size: '1792x1024',
  n: 1,
})

console.log(res.data[0].url)

异步这条路要自己把提交与轮询串起来，Node.js 版：

javascript
const BASE = 'https://open.pikpikgo.com/v1'
const KEY = process.env.PIKPIK_API_KEY

async function call(path, body) {
  const res = await fetch(BASE + path, {
    method: body ? 'POST' : 'GET',
    headers: {
      Authorization: `Bearer ${KEY}`,
      'Content-Type': 'application/json',
    },
    body: body ? JSON.stringify(body) : undefined,
  })
  const json = await res.json()
  if (json.error) throw new Error(json.error.message)
  return json
}

// 提交出图并等待完成：轮询间隔 3 秒，最多等 10 分钟
async function generateImage(prompt) {
  const job = await call('/images/generations', { prompt, size: '16:9', async: true })
  const id = job.tasks[0].id

  for (let i = 0; i < 200; i++) {
    const task = await call(`/tasks/${id}`)
    if (task.status === 'succeeded') return task.data[0].url
    if (task.status === 'failed') throw new Error(task.error || '生成失败')
    await new Promise((r) => setTimeout(r, 3000))
  }
  throw new Error('等待超时')
}

generateImage('赛博朋克风格的雨夜街道，霓虹倒影').then(console.log)

Python 版（这里演示的是出视频，轮询间隔放宽到 5 秒）：

python
import os, time, requests

BASE = "https://open.pikpikgo.com/v1"
KEY = os.environ["PIKPIK_API_KEY"]
HEAD = {"Authorization": f"Bearer {KEY}", "Content-Type": "application/json"}


def call(path, body=None):
    if body is None:
        r = requests.get(BASE + path, headers=HEAD, timeout=30)
    else:
        r = requests.post(BASE + path, headers=HEAD, json=body, timeout=60)
    data = r.json()
    if "error" in data:
        raise RuntimeError(data["error"]["message"])
    return data


def generate_video(prompt, seconds=5):
    job = call("/video/generations", {"prompt": prompt, "duration": seconds})

    # 出片通常 1~5 分钟，轮询间隔 5 秒即可，间隔过密只会白白消耗频控额度
    for _ in range(200):
        task = call(f"/video/generations/{job['task_id']}")
        if task["status"] == "succeeded":
            return task["url"]
        if task["status"] == "failed":
            raise RuntimeError(task["error"] or "生成失败")
        time.sleep(5)
    raise TimeoutError("等待超时")


print(generate_video("镜头缓慢推进，少女回头微笑"))