Eigen Banana Image Edit Studio
Upload or link any photo and edit it by inputing your desired changes.
Checking backend...
Source Image
Drag a file or paste a URL.
Drag & Drop or tap to choose
No file selected
Filled automatically after upload; you can override it with any public URL.
Prompt
Results are returned as a shareable URL.
Advanced Settings
Choose the format for the generated image saved to S3.
Path where the output image will be stored. Generated automatically each run.
🔌 API Access
Invoke Eigen Banana through our standard /api/v1/generate endpoint.
🔑 API Keys
Use your Eigen AI API key for authentication. Manage your API keys →
POST
/api/v1/generate
Generate Images
Upload a file or point to an image URL. The service stores the result on S3 and returns a download URL.
Cost:
0.025 credits per image
Request (cURL)
curl -X POST https://app.eigenai.com/api/v1/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-F "model=eigen-banana-qwen-image-edit" \
-F "prompt=Transform this image into a clean cartoon style. Do not add people if the original has none. Do not add logos." \
-F "image_path=client/image.png" \
-F "output_format=png" \
-F "num_inference_steps=4" \
-F "true_cfg_scale=0.5" \
-F "downsizing_mp=0" \
-F "binary_response=false" \
-F "s3_output_path=s3://chatbot-images-eigenai/banana_$(date +%s)_$RANDOM.png" \
| jq -r '.s3_url'
Request (Python)
import requests
url = "https://app.eigenai.com/api/v1/generate"
headers = {"Authorization": "Bearer YOUR_API_KEY"}
with open("image.png", "rb") as source:
files = {"image_file": source}
data = {
"model": "eigen-banana-qwen-image-edit",
"image_path": "client/image.png",
"prompt": "Transform this image into a clean cartoon style. Do not add people if the original has none. Do not add logos.",
"output_format": "png",
"num_inference_steps": "4",
"true_cfg_scale": "0.5",
"downsizing_mp": "0",
"binary_response": "false",
"s3_output_path": "s3://chatbot-images-eigenai/banana_example.png"
}
response = requests.post(url, headers=headers, data=data, files=files)
response.raise_for_status()
result = response.json()
print("S3 URL:", result.get("s3_url"))
Request (Node.js)
import axios from "axios";
import FormData from "form-data";
import fs from "fs";
const form = new FormData();
form.append("model", "eigen-banana-qwen-image-edit");
form.append("prompt", "Transform this image into a clean cartoon style. Do not add people if the original has none. Do not add logos.");
form.append("image_file", fs.createReadStream("image.png"));
form.append("image_path", "client/image.png");
form.append("output_format", "png");
form.append("num_inference_steps", "4");
form.append("true_cfg_scale", "0.5");
form.append("downsizing_mp", "0");
form.append("binary_response", "false");
form.append("s3_output_path", "s3://chatbot-images-eigenai/banana_example.png");
const response = await axios.post("https://app.eigenai.com/api/v1/generate", form, {
headers: {
...form.getHeaders(),
Authorization: "Bearer YOUR_API_KEY"
}
});
console.log("S3 URL:", response.data?.s3_url);
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
model |
string | ✅ Yes | Must be eigen-banana-qwen-image-edit. |
prompt |
string | ✅ Yes | Instruction describing the transformation. |
image_file |
file | One of ✅ | Upload the source image (PNG, JPG, WEBP). |
image_path |
string | One of ✅ | Remote URL or reference path (auto-filled when using the Studio uploader). |
binary_response |
boolean | ✅ Yes | Set to false so the service returns an S3 URL instead of binary data. |
output_format |
string | ✅ Yes | png or jpg output stored on S3. |
num_inference_steps |
number | ✅ Yes | Number of refinement steps (higher values improve fidelity). |
true_cfg_scale |
number | ✅ Yes | Guidance scale controlling how strongly the prompt is applied. |
downsizing_mp |
number | ✅ Yes | Leave at 0 to keep original resolution. |
s3_output_path |
string | ✅ Yes | Destination bucket/key for the output image (e.g. s3://chatbot-images-eigenai/banana_example.png). |
Output
Model ready
Upload an image or paste a URL, adjust the prompt, then hit generate.
Cost: 0.025 credits per image.