Chenyu AIGC - AI Video & Image Generation

Generate videos and images using AI models through the Chenyu Studio AIGC orchestration API.

When to Use

• User wants to generate a video from text prompt
• User wants to generate a video from an image (first/last frame)
• User wants to extend or remix a video
• User wants to generate AI images
• User wants to check status of a generation task
• User wants to list available AI models

When NOT to Use

• User wants to analyze or understand existing videos (use video-analysis skill)
• User wants to download videos from social platforms (use video-fetch skill)
• User wants to manage digital humans or clone voices (use chenyu-core skill)

Authentication

Authorization: Bearer $CHENYU_API_KEY

Base URL: $CHENYU_BASE_URL (default: https://chenyu.pro)

Workflow

1. Discover recipes — list available AI models (see below)
2. Get recipe schema — check what inputs/parameters the recipe accepts
3. Execute — submit a generation task → see execute-recipe.md
4. Poll & manage — track status, get output, cancel → see manage-tasks.md

Step 1: List Available Recipes

curl -s "$CHENYU_BASE_URL/api/v1/aigc/recipes" \
  -H "Authorization: Bearer $CHENYU_API_KEY" | jq '.data[] | {recipe_id, name, slug, description, output_type}'

Each recipe represents an AI model capability. Key response fields:

• recipe_id — use this ID when executing
• slug — human-readable identifier (e.g. volcengine-seedance-v1-pro)
• output_type — what it produces: video, image, audio

Step 2: Get Recipe Schema

curl -s "$CHENYU_BASE_URL/api/v1/aigc/recipes/{recipe_id}/schema" \
  -H "Authorization: Bearer $CHENYU_API_KEY" | jq '.data'

The schema tells you:

• typed_inputs_schema.definitions — accepted input types and their fields
• parameters_schema — available parameters with constraints (min/max/enum)
• credit_cost / credit_cost_rules — how many credits it costs

After getting the schema, read execute-recipe.md for execution details.