infinimo-ai-design-image-create

AI Image Generation

About Infinimo AI Design

Infinimo AI Design is an AI design platform for cross-border e-commerce product images and video. Built for Amazon, Shopify, and global sellers, it delivers studio-quality white-background heroes, lifestyle scenes, and listing-ready visuals. This skill covers multi-style image generation and editing so teams can produce commercial, marketplace-ready assets quickly.

This skill calls the Infinimo AI Design image generation API (same as the web General image generation page): model selection, reference upload, job submission, and result retrieval.

Web page: https://design.infinimo.ai/?source=q-i-d-clawhub

Authentication & base URL

• Base URL: https://www.clawec.com/api
• Token: Sign up and sign in at https://design.infinimo.ai/?source=q-i-d-clawhub; or use an API Key from https://design.infinimo.ai/api-key?source=q-i-d-clawhub as the Token
• Headers:
  • Token: (session token or API Key)
  • Content-Type: application/json (JSON endpoints)
  • Time-Zone: Asia/Shanghai (optional; matches the web app)

Read INFINIMO_TOKEN or INFINIMO_API_KEY from the environment first; if unset, ask the user—never hardcode credentials.

Note: Image generation uses the Token header, not Authorization: Bearer used by /aigc/tool/* open tools.

Common parameters

The web app attaches these on every request (query for GET, body fields for POST):

Scripts may default to: platform=1&terminal=4&language=en (same for upload).

---

End-to-end flow

1. GET  /aigc/ec_media/image/create/dic     → models / ratios / sizes
2. POST /upload/image                       → upload reference images (optional, max 12)
3. POST /aigc/ec_media/image/create         → submit generation job
4. GET  /aigc/ec_media/image/create/logs    → poll for results

Generation is usually async. After a successful submit, poll logs until urls is non-empty or fail=true.

---

1. Get model dictionary

GET /aigc/ec_media/image/create/dic

Returns models, aspect ratios, and resolutions. Use each option’s id when submitting a job.

Request

curl -s -G "https://www.clawec.com/api/aigc/ec_media/image/create/dic" \
  -H "Token: $INFINIMO_TOKEN" \
  --data-urlencode "platform=1" \
  --data-urlencode "terminal=4" \
  --data-urlencode "language=en"

Or use the script:

bash scripts/dic.sh

Response data

models items (TypeRespString + optional fields)

ratios / sizes items

---

2. Upload reference images

POST /upload/image

Upload local images via multipart/form-data; use returned URLs in the images array on create.

Limits (same as web):

• image/* only
• Up to 12 reference images per job
• Response data.url or data.path; prepend Base URL if path is relative

Request

curl -s -X POST "https://www.clawec.com/api/upload/image" \
  -H "Token: $INFINIMO_TOKEN" \
  -F "file=@/path/to/reference.jpg" \
  -F "platform=1" \
  -F "terminal=4" \
  -F "language=en"

Or use the script:

bash scripts/upload.sh /path/to/reference.jpg

Example data

{
  "url": "https://cdn.example.com/xxx.jpg"
}

---

3. Submit image generation

POST /aigc/ec_media/image/create

Reference images in prompts: after upload, refer to them in order as [Image 1], [Image 2], … e.g. Place the product from [Image 1] on a pure white background.

Request

curl -s -X POST "https://www.clawec.com/api/aigc/ec_media/image/create" \
  -H "Token: $INFINIMO_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A cute orange cat, product photography style",
    "model": "MODEL_ID",
    "ratio": "RATIO_ID",
    "size": "SIZE_ID",
    "images": ["https://cdn.example.com/ref.jpg"],
    "platform": 1,
    "terminal": 4,
    "language": "en"
  }'

Or use the script:

bash scripts/create.sh \
  --prompt "A cute orange cat" \
  --model MODEL_ID \
  --ratio RATIO_ID \
  --size SIZE_ID \
  --images '["https://cdn.example.com/ref.jpg"]'

Success returns status: 1; data is usually {}. Fetch output URLs from logs.

---

4. Get generation results

GET /aigc/ec_media/image/create/logs

Request

curl -s -G "https://www.clawec.com/api/aigc/ec_media/image/create/logs" \
  -H "Token: $INFINIMO_TOKEN" \
  --data-urlencode "start=1" \
  --data-urlencode "size=5" \
  --data-urlencode "platform=1" \
  --data-urlencode "terminal=4" \
  --data-urlencode "language=en"

Or use the script:

bash scripts/logs.sh 1 5

Response data

items (ImageResultV2)

Polling

1. After submit, call logs?start=1&size=5
2. Take the newest row (or match by prompt); check urls and fail
3. If urls is empty and fail is not true, wait 3–5s and retry
4. The web app also listens on WebSocket wss://www.clawec.com/api/aigc/socket for image_result_refresh; scripts can poll only

Delete record (optional)

GET /aigc/ec_media/image/log/delete?id=

---

Response envelope

All endpoints share this wrapper:

{
  "status": 1,
  "code": 200,
  "msg": "success",
  "data": { ... },
  "pointInfo": { "type": 0, "point": 0 }
}

• status: 1 = success, 0 = failure
• code: 2001 = not signed in / invalid Token, 2002 = insufficient credits
• data: payload
• pointInfo: credit usage

See references/response-schema.md for full field notes.

---

Workflow

1. Confirm prompt and whether reference images are needed
2. Ensure INFINIMO_TOKEN or INFINIMO_API_KEY is set
3. Call dic for model / ratio / size; pick per user request or first defaults
4. If references exist, upload each file for URLs
5. create to submit the job
6. Poll logs until urls appear or fail is confirmed
7. Return image links and params; on failure, explain and suggest prompt tweaks or top-up

Output format

Default to an English summary with:

• Model, ratio, resolution (titles + ids)
• Prompt and reference count
• Result URLs; on failure, note fail
• Credits used if pointInfo or model point is present

Example

Input: Generate a wireless earbuds product shot on white background using my uploaded photo

Steps:

1. dic.sh → model=flux-pro, ratio=1:1, size=2K
2. upload.sh ref.jpg → https://cdn.../ref.jpg
3. create.sh --prompt "White background product photo, wireless earbuds, soft studio lighting" --model flux-pro --ratio 1:1 --size 2K --images '["https://cdn.../ref.jpg"]'
4. Poll logs.sh → urls: ["https://cdn.../output.png"]

Summary:

• Model: Flux Pro | Ratio 1:1 | Size 2K
• Prompt: White background product photo, wireless earbuds, soft studio lighting
• References: 1 image
• Result: https://cdn.../output.png