0crUpgrade

Image generation API

Image Generation

Generate images using 190+ AI models from a single endpoint. All image models use POST /api/generate with different modelKey values.

Endpoint

POST /api/generate

Request body

ParameterTypeRequiredDescription
modelKeystringModel identifier (see model families below)
promptstringText description of the image to generate
aspectRatiostringAspect ratio: "1:1", "16:9", "9:16", "4:3", "3:4"
inputUrlstringURL of input image (for image-to-image / edit models)
styleRefstring | string[]Style reference image URL(s)
charRefstring | string[]Character reference image URL(s)
refImageUrlsstring[]General reference images
imageQualitystring"Low", "Medium", "High"
qualitystring"HD", "1080p", "4K"
brandIdstringBrand DNA ID to bake brand guidelines into the prompt

Response

{
  "success": true,
  "predictionId": "pred_abc123...",
  "status": "processing",
  "model": "openai/gpt-image-2/text-to-image",
  "creditCost": 5,
  "estimatedTime": "~15s"
}

Quick example

curl -X POST https://api.oyefilmy.com/api/generate \
  -H "Authorization: Bearer of_live_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "modelKey": "gpt",
    "prompt": "A photorealistic portrait of a woman with golden hour lighting, shallow depth of field"
  }'

Model families

Choose a model family to see all available model keys and their capabilities:

auto_awesome

GPT Image

OpenAI GPT Image 1/1.5/2 — text-to-image and image editing

gpt, gpt_2, gpt5, gpt_4o, ...

auto_awesome

Seedream

ByteDance Seedream v4/v4.5/v5 — sequential and edit modes

seedream_5_pro_t2i, seedream_5_pro_edit, seedream_v5_lite, seedream_v4_5, ...

auto_awesome

Flux

Black Forest Labs FLUX.1/2 — Pro, Flex, Klein, Kontext, LoRA

flux_max, flux_pro, flux_flex, flux_kont_max, ...

auto_awesome

Imagen / Google

Google Imagen 3/4 and Gemini native image generation

imagen, imagen_2, imagen_4_fast, imagen_4, ...

auto_awesome

Wan Image

Alibaba Wan 2.5/2.6/2.7 — text-to-image and image editing

wan_2_7_t2i, wan_2_7_edit, wan_2_7_pro_t2i, wan_2_6_t2i, ...

auto_awesome

Qwen Image

Alibaba Qwen Image — text-to-image and edit modes

qwen_image_2_0_t2i, qwen_image_2_0_edit, qwen_image_t2i_max, qwen_image_t2i_plus

auto_awesome

Grok Imagine

xAI Grok Imagine Image — quality and standard variants

grok_imagine_image_quality_t2i, grok_imagine_image_quality_edit, grok_imagine_image_t2i

auto_awesome

Kling Image

Kling Image 4.0: up to 9 pictures a run, a related set of 2 to 10, or one long product page or webpage

kling_image_4_t2i, kling_image_4_set, kling_image_4_long

auto_awesome

Tencent Vega

Tencent WAND Vega Image 1.5 Pro: text-to-image and edit at 1K, 2K or 4K

vega_1_5_pro_t2i, vega_1_5_pro_edit

auto_awesome

Microsoft MAI

Microsoft MAI Image 2.5 — Flash and Standard variants

mai_2_5_flash_t2i, mai_2_5_edit, mai_2_5_t2i, mai_2_5_flash_edit

auto_awesome

Nano Banana

Google Nano Banana — Lite, Standard, Pro with edit and reference modes

banana_pro_t2i_ultra, banana_pro_edit_ultra, banana_2_lite_t2i, banana_2_ref2img

auto_awesome

Youchuan

Youchuan v8.1 — text-to-image, image-to-image, style transfer, background removal

youchuan_v8_1_t2i, youchuan_v8_1_i2i, youchuan_v8_1_style_transfer, youchuan_v8_1_remove_bg

Poll for results

GET /api/generate/status/:predictionId

# Success response:
{
  "status": "completed",
  "outputUrl": "https://oyefilmy-assets.s3.amazonaws.com/.../output.png",
  "outputType": "image",
  "completedAt": "2026-07-27T14:00:00.000Z"
}