Uwear Agent Skills
Build with Uwear from one agent-readable page.
Download UWEAR.md, inspect the manifest, and give your coding agent the API rules for fashion product photos, try-on style images, edits, upscales, and video.
curl -L https://uwear.ai/UWEAR.md -o UWEAR.md
# Load UWEAR.md as context.
# Keep the API key server-side.
export UWEAR_API_KEY="..."
# Core contract
POST https://api.uwear.ai/generation
GET https://api.uwear.ai/generation/{generation_id}Base URL
https://api.uwear.ai
Auth
Authorization: Bearer YOUR_API_KEY
Field style
snake_case for requests and responses
Create job
POST /generation
Resource Index
Crawl this first.
These are the durable URLs an agent should inspect before calling the API or writing integration code.
| Resource | Type | Use |
|---|---|---|
| Agent skills hub https://uwear.ai/skills | HTML | Canonical starting point for agents and crawlers. |
| UWEAR.md https://uwear.ai/UWEAR.md | Markdown | Load this file as task context before building with the Uwear API. |
| Skills manifest https://uwear.ai/skills/manifest.json | JSON | Machine-readable index of Uwear agent resources and API rules. |
| llms.txt https://uwear.ai/llms.txt | Text | Broad crawler summary of Uwear products, API, and agent references. |
| API docs https://platform.uwear.ai/api/introduction | Docs | Human-facing API documentation for Uwear generation routes. |
| Agent API guide https://uwear.ai/blog/ai-fashion-generation-api-for-agents | Guide | Long-form explanation of the agent-ready generation workflow. |
Core Contract
Use one async generation model.
Uwear generation, edit, upscale, and video workflows share the same mental model: create a job, wait for a terminal state, then persist the generated assets.
Agent Workflow
The path from prompt to production assets.
1. Load the skill
Fetch UWEAR.md and keep it in the agent context before writing code or planning a workflow.
2. Pick the asset path
Use clothing items for garments, avatars for reusable models, and generation results for edits, upscales, or video follow-ups.
3. Create one job per output
Call POST /generation with explicit model_slug, use_case, camera, aspect_ratio, resolution, and num_images settings.
4. Wait safely
Poll for scripts or use signed webhooks for production and catalog-scale workflows.
5. Persist results
Download completed assets to your own storage before the temporary delivery URLs expire.
Skill Coverage
What UWEAR.md teaches the agent.
The markdown file is the practical contract. It tells an agent how to use documented routes, handle temporary URLs, avoid leaking API keys, and structure real catalog or try-on work.
- Clothing item creation and reuse
- Async generation for images, edits, upscales, and video
- Polling and signed webhook completion patterns
- Generation result retrieval and semantic search
- Saved avatars for consistent models
- Bulk catalog manifests and retry-safe processing
- Storefront try-on backend architecture
- Agent guardrails for secrets, client code, and temporary URLs
Minimal Request Shape
Start with explicit settings.
Agents should prefer structured API fields over vague prompts. The prompt describes the shoot; the API fields define the workflow.
POST /generation
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
{
"clothing_item_id": 123,
"model_slug": "seedream-4-5",
"use_case": "generate",
"prompt": "professional ecommerce studio photo",
"num_images": 2,
"camera": "full_body_front",
"aspect_ratio": "9:16",
"resolution": "2K",
"enhance_user_prompt": true
}Agent starting point: fetch the skill, then build.
For production, keep the Uwear API key on the backend, process large catalogs through a queue, and store every generated result before the 4-hour URL window closes.