跳转到主要内容

​
接口说明

为了方便开发者调用不同的图像生成模型，Aihubmix 提供了统一的图形接口，支持以下多种主流模型： 接口规格：
Curl
curl https://aihubmix.com/v1/models/<model_path>/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "your prompt"
  }
}'
其中：
  • https://aihubmix.com/v1/models/ 为固定前缀
  • <model_path> 为模型路径，支持：
    • opanai/gpt-image-1
    • opanai/dall-e-3
    • imagen-4.0-ultra-generate-001
    • imagen-4.0-generate-001
    • imagen-4.0-fast-generate-001
    • imagen-4.0-fast-generate-preview-06-06
    • imagen-3.0-generate-002
    • flux-kontext-max
    • flux-kontext-pro
    • FLUX.1-Kontext-pro
    • FLUX-1.1-pro
    • Qwen-Image
    • Qwen-Image-Edit
    • doubao-seedream-4-0-250828
    • ideogram/V3
    • stability/Stable-Diffusion-3-5-Large
    • irag-1.0
    • ernie-irag-edit
  • sk-*** 为你在 Aihubmix 生成的 API Key
下面是各个模型的调用方法。

​
OpenAI

​
支持的模型：

  • gpt-image-1
  • dall-e-3
不支持 2 个逆向模型 gpt-4o-image-vip 和 gpt-4o-image

​
调用方法：

curl https://aihubmix.com/v1/models/opanai/gpt-image-1/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "A deer drinking in the lake, Sakura petals falling, green and clean water, japanese temple, dappled sunlight, cinematic lighting, expansive view, peace",
    "size": "1024x1024", 
    "n": 1,
    "quality": "high",
    "moderation": "low",
    "background": "auto"
  }
}'

​
Google Imagen

​
支持的模型：

  • imagen-4.0-ultra-generate-001
  • imagen-4.0-generate-001
  • imagen-4.0-fast-generate-001
  • imagen-4.0-fast-generate-preview-06-06
  • imagen-3.0-generate-002
不支持 gemini 系列

​
调用方法：

curl https://aihubmix.com/v1/models/google/imagen-4.0-fast-generate-001/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "A deer drinking in the lake, Sakura petals falling, green and clean water, japanese temple, dappled sunlight, cinematic lighting, expansive view, peace",
    "numberOfImages": 1
  }
}'

​
Flux

​
支持的模型：

  • flux-kontext-max
  • flux-kontext-pro
  • FLUX.1-Kontext-pro
  • FLUX-1.1-pro
  • flux-kontext-max 和 flux-kontext-pro 为异步端口，需要分 2 步请求。该系列速度极快，几乎秒出。
  • FLUX.1-Kontext-pro 和 FLUX-1.1-pro 使用通用接口，一步到位
  • flux-Kontext-max 支持文生图和参考图生图，flux-kontext-pro 端口只支持文生图
  • 画幅比例 aspect_ratio 支持的范围是：21:9 到 9:21
  • 审核宽松度参数 safety_tolerance 范围在 0 到 6 之间0 表示最严格6 表示最宽松。使用参考图时限制为 2。
  • 参考图支持 URL 或 Base64 格式，推荐采用前者
  • 更多细节参考：FLUX API Reference

​
调用方法：

FLUX 一步绘图： 支持 FLUX.1-Kontext-pro 和 FLUX-1.1-pro 
curl 'https://aihubmix.com/v1/images/generations' \
  -H 'accept: */*' \
  -H 'accept-language: zh-CN' \
  -H 'authorization: Bearer sk-***' \
  -H 'content-type: application/json' \
  --data-raw '{"prompt":"a cat in the garden, cute cartoon","model":"FLUX.1-Kontext-pro","safety_tolerance":6}' \
| jq -r '.data[0].b64_json' \
| base64 -D > "$HOME/Desktop/image_extract.png"
flux 异步类型第一步：发出绘图请求 
curl https://aihubmix.com/v1/models/bfl/flux-kontext-max/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "A deer drinking in the lake, Sakura petals falling, green and clean water, japanese temple, dappled sunlight, cinematic lighting, expansive view, peace. reference style as the given image:",
    "aspect_ratio": "2:3",
    "safety_tolerance": 6
  }
}'
输出结果示例： 
{"output":[{"polling_url":"https://api.us1.bfl.ai/v1/get_result?id=f9821dbf-f134-41c2-aaa8-5405fb665c76","taskId":"api.us1.***"}]}%
flux 异步类型第二步：获取生成的图片链接 tasks/ 后面的字符串替换成第一步获取的 taskId：
获取图片
curl https://api.aihubmix.com/v1/tasks/api.us1.bfl.*** \
    -H "Authorization: Bearer sk-***" \
    -X GET

​
Qwen

​
支持的模型：

  • Qwen-Image
  • Qwen-Image-Edit

​
调用方法：

curl https://aihubmix.com/v1/models/qianfan/qwen-image/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
  -d '{
    "input": {
      "prompt": "A beautiful sunset over a calm ocean",
      "refer_image": "",
      "n": 1,
      "size": "1024x1024",
      "guidance": 7.5,
      "watermark": false
    }
  }'

​
doubao-seedream-4-0-250828

​
支持的模型：

  • doubao-seedream-4-0-250828

​
调用方法：

curl https://aihubmix.com/v1/models/doubao/doubao-seedream-4-0-250828/predictions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-***" \
  -d '{
    "input": {
      "prompt": "星际穿越，黑洞，黑洞里冲出一辆快支离破碎的复古列车，抢视觉冲击力，电影大片，末日既视感，动感，对比色，oc渲染，光线追踪，动态模糊，景深，超现实主义，深蓝，画面通过细腻的丰富的色彩层次塑造主体与场景，质感真实，暗黑风背景的光影效果营造出氛围，整体兼具艺术幻想感，夸张的广角透视效果，耀光，反 射，极致的光影，强引力，吞噬",
      "size": "2K",
      "sequential_image_generation": "disabled",
      "stream": false,
      "response_format": "url",
      "watermark": true
    }
  }'
输出结果示例： 
{"output":[{"url":"https://ark-content-generation-v2-cn-beijing.tos-cn-beijing.volces.com/doubao-seedream-4-0/021758***"}]}

​
Ideogram V3

​
支持的模型：

  • V3
  1. V3 以下版本（V_2、V_1 等）为旧接口，暂不支持
  2. 返回的链接需要打开代理网络才能获取

​
调用方法：

curl https://aihubmix.com/v1/models/ideogram/V3/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "A deer drinking in the lake, Sakura petals falling, green and clean water, japanese temple, dappled sunlight, cinematic lighting, expansive view, peace, in the style of Pixar 3D",
    "rendering_speed": "QUALITY",
    "aspect_ratio": "2x1"
  }
}'

​
Stability

​
支持的模型：

  • Stable-Diffusion-3-5-Large

​
调用方法：

curl https://aihubmix.com/v1/models/stability/Stable-Diffusion-3-5-Large/predictions \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
    -d '{
  "input": {
    "prompt": "A deer drinking in the lake, Sakura petals falling, green and clean water, japanese temple, dappled sunlight, cinematic lighting, expansive view, peace",
    "n": 1
  }
}'

​
百度系列

​
支持的模型：

  • ERNIE iRAG
  • ERNIE iRAG Edit

​
调用方法：

curl https://aihubmix.com/v1/models/qianfan/irag-1.0/predictions\
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer sk-***" \
  -d '{
    "input": {
      "prompt": "A beautiful sunset over a calm ocean",
      "refer_image": "",
      "n": 1,
      "size": "1024x1024",
      "guidance": 7.5,
      "watermark": false
    }
  }'