{"id":51026385,"url":"https://github.com/siliconflow/bizyair-skill","last_synced_at":"2026-06-21T20:01:22.551Z","repository":{"id":359499439,"uuid":"1246292694","full_name":"siliconflow/bizyair-skill","owner":"siliconflow","description":"连接 bizyair.cn 云端生态，赋予 Agent 图像/视频生成、AI 应用执行与 ModelZoo 模型调用等 AIGC 能力，助力实现更高效的视觉创作与 AI 任务流自动化。","archived":false,"fork":false,"pushed_at":"2026-05-22T08:44:53.000Z","size":111,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-22T14:30:45.752Z","etag":null,"topics":["agent-skill","agent-skills","ai-agents","aigc","bizyair","text-to-image"],"latest_commit_sha":null,"homepage":"https://bizyair.cn","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/siliconflow.png","metadata":{"files":{"readme":"README.en.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-22T03:55:47.000Z","updated_at":"2026-05-22T08:44:57.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/siliconflow/bizyair-skill","commit_stats":null,"previous_names":["siliconflow/bizyair-skill"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/siliconflow/bizyair-skill","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/siliconflow%2Fbizyair-skill","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/siliconflow%2Fbizyair-skill/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/siliconflow%2Fbizyair-skill/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/siliconflow%2Fbizyair-skill/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/siliconflow","download_url":"https://codeload.github.com/siliconflow/bizyair-skill/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/siliconflow%2Fbizyair-skill/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":34623906,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-06-21T02:00:05.568Z","response_time":54,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent-skill","agent-skills","ai-agents","aigc","bizyair","text-to-image"],"created_at":"2026-06-21T20:01:21.825Z","updated_at":"2026-06-21T20:01:22.542Z","avatar_url":"https://github.com/siliconflow.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\n  \u003ch1\u003eBizyAir Skill\u003c/h1\u003e\n\n\n\u003cdiv align=\"center\"\u003e\n  \u003ca href=\"https://bizyair.cn\" target=\"_blank\" rel=\"noopener noreferrer\"\u003e\n    \u003cimg\n      src=\"https://plugin.bizyair.com/public/logo_lightmode.png\"\n      alt=\"BizyAir\"\n      width=\"370\"\n    /\u003e\n\n  \u003c/a\u003e\n\n---\n\u003cdiv align=\"center\"\u003e\n\n\u003e BizyAir Skill is an AIGC skill pack designed for mainstream AI agents. Powered by BizyAir cloud compute, it removes complex setup and parameter tuning.\u003cbr\u003e**With natural language alone, go from idea to finished image and video output in one smooth flow, giving your AI agent professional-grade AIGC capability.**\n\n\u003c/div\u003e\n\n\n\u003ca href=\"https://skills.sh\"\u003e\u003cimg src=\"https://img.shields.io/badge/skills.sh-Compatible-green?style=for-the-badge\" alt=\"Skills\"\u003e\u003c/a\u003e\n\u003ca href=\"https://agentskills.io\"\u003e\u003cimg src=\"https://img.shields.io/badge/AgentSkills-Standard-blue?style=for-the-badge\" alt=\"AgentSkills\"\u003e\u003c/a\u003e\n\u003ca href=\"https://claude.ai/code\"\u003e\u003cimg src=\"https://img.shields.io/badge/Claude%20Code-Skill-blueviolet?style=for-the-badge\" alt=\"Claude Code\"\u003e\u003c/a\u003e\n\n\n[Installation](#-installation) · [Usage](#-how-to-talk-to-it) · [Examples](#-how-it-interacts-with-you) · [CLI Commands](#-developer--debug-commands)\n\n\n\n**[中文](./README.md) · [English](./README.en.md)**\n\u003c/div\u003e\n\n****\n\n\n\n## ✨ What it can do\n\n| Core capability | Typical scenarios | Interaction style |\n| :--- | :--- | :--- |\n| **Image Generation** | Text-to-image, image-to-image, editing, outpainting | **5 curated models** + ModelZoo search + AI app search |\n| **Video Generation** | Text-to-video, image-to-video, lip sync, etc. | **5 curated models** + ModelZoo search + AI app search |\n| **AI Apps \u0026 ModelZoo** | Inspect and run BizyAir apps and underlying ModelZoo endpoints | Direct links, object inspection, dual-track search |\n| **Smart Parameter Engine** | Intent completion, prompt refinement | \"Prefill confirmation card\" instead of blank forms |\n\n---\n\n## ⭐️ Prerequisites\n\n- **Python 3.9+**\n- A BizyAir API key ready (sign in, hover over your avatar in the top-right corner, then click [\"API Keys\"](https://bizyair.cn/user/api-key) to create one)\n- Sufficient BizyAir balance in your account\n\n---\n\n## 🚀 Installation\n\nSend this in your agent chat:\n\n\u003e Help me install the BizyAir skill from` https://github.com/siliconflow/bizyair-skill `\n\n\nOr run directly:\n\n```bash\ngh skill install siliconflow/bizyair-skill\n```\n\n---\n\n### Config file notes (How to configure the API key)\n\n**Option A: use an environment variable**\n\n```bash\nexport BIZYAIR_API_KEY=\"your BizyAir API Key\"\n```\n\n**Option B: write it into the config file**\n\nThe `config.json` in the skill root stores the BizyAir API key and batch-related settings.\n\nYou can edit it manually, or just give the API key to the assistant and let it write the file for you. **(Recommended)**\n\n```json\n{\n  \"credentials\": {\n    \"api_key\": \"Put your BizyAir API Key here\"\n  },\n  \"batch\": {\n    \"max_concurrency\": 3,\n    \"max_tasks\": 5\n  }\n}\n```\n\n\u003e **Notes**: `max_concurrency` — maximum concurrent tasks, default 3, hard cap 5; `max_tasks` — maximum tasks in a single batch, default 5, hard cap 10.\n\n---\n\n## 💬 How to talk to it\n\nOnce configured, just talk to the assistant in natural language (pattern: \"Use bizyair to ...\"):\n\n**🎨 Image tasks:**\n* \"Use bizyair to create a warm brown coffee advertising poster.\"\n* \"Use bizyair to make an elegant anime-style avatar.\"\n* \"I want to generate an image. Show me the common BizyAir menu first.\"\n\n**🎬 Video tasks:**\n* \"Use bizyair to make a 10-second rainy neon street short video.\"\n* \"Animate this portrait image with bizyair.\"\n* \"Use bizyair to turn this portrait and this audio clip into a lip-sync video.\"\n\n**🧩 AI App / Workflow tasks:**\n* \"Run this BizyAir app: `https://bizyair.cn/...`\"\n* \"Help me understand what this BizyAir workflow ID does and what assets it needs.\"\n* \"Find a few BizyAir workflows suitable for e-commerce key visual retouching.\"\n\n**🛠️ Status and utility tasks:**\n* \"Check my BizyAir API key.\"\n* \"Check my BizyAir balance.\"\n\n---\n\n## 🤖 How it interacts with you\n\n### 1. Curated video / image menus\nWhen your request is vague (e.g. \"help me make a video\"), the assistant shows a curated model menu first.\n\nFor example, you can say:\n\u003e \"I want to make a video. Show me the common BizyAir video entry points first.\"\n\nThe assistant will respond like:\n\u003e \"Pick a model 👇\n\u003e\n\u003e 【⭐ Curated · 5 common video models】\n\u003e 1. 🎬 V.3.1.Pro: cinematic mood, lens feel, stronger visual impact.\n\u003e 2. 🐎 HappyHorse: fast output, comfortable price-performance.\n\u003e 3. 🐉 Kling 3.0 Pro: stronger for action, movement, and impact, native 4K.\n\u003e 4. 🌊 Wanxiang 2.7: strong Chinese understanding, audio sync, multi-shot editing.\n\u003e 5. 💃 Seedance 2.0: character motion, dance, and role consistency.\n\u003e\n\u003e 【🧪 ModelZoo Search · ✨ NEW】\n\u003e 6. Direct calls to underlying model APIs — explicit params, predictable per-call billing.\n\u003e\n\u003e 【🔎 AI App Search】\n\u003e 7. Find ready-made workflow templates for complex multi-step pipelines.\n\u003e\n\u003e I'd suggest starting with option 4 — your request hasn't locked down a specific style yet, so Wanxiang 2.7 tends to be more stable.\n\u003e Just reply 1–7; if you already know the model name or number, we can use that directly.\"\n\n### 2. Link diagnosis\nIf you send an unfamiliar App link, the assistant will fetch the underlying inspection result (`info`) and explain in natural language:\n\n* What the workflow does.\n* What input assets it needs (image / audio / text).\n* Whether it can be run directly or requires the web interface.\n\nFor example, you can say:\n\u003e \"Help me inspect this BizyAir app: https://bizyair.cn/community/app/46086\"\n\nThe assistant will respond like:\n\u003e \"All checked. This workflow looks healthy.\n\u003e\n\u003e - **What it does**: it's a **Wai-Illustrious anime text-to-image workflow**, with HD upscaling and hand/face repair.\n\u003e - **What you need to provide**: no base image needed, just a solid prompt.\n\u003e - **What it outputs**: a complete set of finished images (including HD upscaled versions).\n\u003e - **Execution status**: ✅ the interface is fully reachable, I can run it for you directly.\n\u003e\n\u003e So — do you want to start shaping the prompt now, or should I first lay out the size and ratio parameter card for review?\"\n\n\u003e **Note**: comfy-ui links (`bizyair.cn/comfy-ui?id=xxx`) are not currently supported for direct execution. The assistant will prompt you to find the corresponding AI App version.\n\n### 3. Smart parameter confirmation card\nOnce a model is chosen, the assistant absorbs your natural-language request and generates a **prefill confirmation card**.\n\nFor example, you can say:\n\u003e \"I'll use option 4, O.2. Make a Chinese coffee poster, landscape layout.\"\n\nThe assistant will respond like:\n\u003e \"## 🎨 O.2 Parameter Confirmation\n\u003e\n\u003e **Prompt**: Chinese coffee poster, landscape layout\n\u003e **Ratio**: 3:2 (options: 1:1 / 2:3 / 3:2 / 4:5 / 5:4 / 3:4 / 4:3 / 16:9 / ...)\n\u003e **Resolution**: 2K (options: 1K / 2K / 4K)\n\u003e\n\u003e Once confirmed, reply \"run\" to generate immediately, or just tell me if you want to change the ratio or resolution.\"\n\n### 4. Dual-track search\nBeyond the 5 curated models, the assistant can take two different search routes depending on what you need:\n\n- **Slot 6 — ModelZoo Search**: direct calls to underlying model APIs, with explicit parameters and predictable per-call billing. Good when you already know the rough route (\"I want a Kling video\", \"find me a specific endpoint\").\n- **Slot 7 — AI App Search**: BizyAir's workflow templates that bundle complex multi-step pipelines into one click. Good when you want a pre-made solution (\"a workflow that swaps backgrounds\", \"an e-commerce key-visual template\").\n\n#### Example 1: find a ModelZoo endpoint\n\nFor example, you can say:\n\u003e \"I want to call an underlying video model API directly. Search me a few kling endpoints.\"\n\nThe assistant will respond like:\n\u003e 🎯 **Picked a few video-side ModelZoo endpoints for you**\n\u003e\n\u003e ModelZoo gives you direct API access — explicit params, predictable per-call billing. Here's what looks closest:\n\u003e\n\u003e 1️⃣ **Kling O3.Pro · FLF to Video (channel edition)**\n\u003e\n\u003e - **Summary**: O3-architecture first/last-frame video model, supports first/last-frame constraints with smooth motion control.\n\u003e - **Category**: FLF to Video\n\u003e - **Runnable directly**: ✅ Yes, after parameter confirmation\n\u003e - **endpoint**: `kling-o3-pro-base/flf-to-video`\n\u003e\n\u003e 2️⃣ **Kling O3.Pro · text-to-video (channel edition)**\n\u003e\n\u003e - **Summary**: O3-architecture text-to-video model, sharp subjects, smooth motion, natural lighting.\n\u003e - **Category**: Text to Video\n\u003e - **Runnable directly**: ✅ Yes, after parameter confirmation\n\u003e - **endpoint**: `kling-o3-pro-base/text-to-video`\n\u003e\n\u003e Tell me a number or endpoint name and I'll move on to the parameter card.\n\n#### Example 2: find an AI app / workflow\n\nFor example, you can say:\n\u003e \"I want to generate an architectural render. Can you check what BizyAir workflows would suit it?\"\n\nThe assistant will respond like:\n\u003e 🎯 **Found some BizyAir objects that fit your needs better**\n\u003e\n\u003e I filtered out the weak matches. Here are the ones worth looking at:\n\u003e\n\u003e 1️⃣ **FLUX建筑效果图高清修复**\n\u003e\n\u003e - **Best for**: Flux.1 D based route, suitable as a candidate for this round.\n\u003e - **Runnable directly**: ✅ Yes\n\u003e - **Link**: https://bizyair.cn/community/app/44479\n\u003e - **ID**: `44479`\n\u003e\n\u003e 2️⃣ **建筑效果图放大**\n\u003e\n\u003e - **Best for**: Flux.2 Klein based route, suitable as a candidate for this round.\n\u003e - **Runnable directly**: ✅ Yes\n\u003e - **Link**: https://bizyair.cn/community/app/44396\n\u003e - **ID**: `44396`\n\u003e\n\u003e Which one do you like? Just tell me the number or ID and I'll proceed.\n\n---\n\n## 👨‍💻 Developer \u0026 Debug Commands\n\n\n**Image / Video base flow:**\n```bash\n# Show common image/video route menus\npython3 scripts/cli.py image-menu\npython3 scripts/cli.py video-menu\n\n# Fixed search routes (AI app candidates)\npython3 scripts/cli.py pick-image \"\u003ckeyword\u003e\" --remote\npython3 scripts/cli.py pick-video \"\u003ckeyword\u003e\" --remote\n\n# ModelZoo endpoint candidates (menu slot 6 / v6)\npython3 scripts/cli.py pick-modelzoo-image \"\u003ckeyword\u003e\"\npython3 scripts/cli.py pick-modelzoo-video \"\u003ckeyword\u003e\"\n\n# Generate a prefilled confirmation card (fixed routes)\npython3 scripts/dispatch.py --model \u003cROUTE_ID\u003e --prefill-card --prompt \"\u003crequest\u003e\"\n\n# General BizyAir platform search (apps / workflows / creations / MCP)\npython3 scripts/cli.py search \"\u003ckeyword\u003e\" --remote\n\n# Browse in-site app listings (filterable by type/sort)\npython3 scripts/cli.py browse --remote-source community --modality Application --sort \"Most Used\"\n\n# Multi-task / parallel — generate batch parameter cards first\npython3 scripts/cli.py batch-prefill --model \u003cROUTE_ID\u003e --batch-json '\u003cJSON task array\u003e'\n# batch-json format examples (each item can be a string or object):\n# Short: '[\"promptA\", \"promptB\", \"promptC\"]'\n# Full: '[{\"prompt\":\"promptA\",\"aspect_ratio\":\"16:9\"}, {\"prompt\":\"promptB\"}]'\n# With concurrency: '{\"concurrency\":2,\"tasks\":[\"promptA\",\"promptB\"]}'\n\n# Batch confirmed execution (run after prefill confirmation)\npython3 scripts/cli.py batch-run --model \u003cROUTE_ID\u003e --batch-json '\u003cJSON task array\u003e' --confirm-run\n\n# Execute a fixed route after explicit confirmation (single task)\npython3 scripts/dispatch.py --model \u003cROUTE_ID\u003e --prompt \"\u003cprompt\u003e\" [other params] --confirm-run\n```\n\n**AI App / Workflow advanced flow:**\n```bash\n# Inspect object details and execution support from a link or ID\npython3 scripts/cli.py info \"\u003clink-or-id\u003e\"\n\n# Generate a prefilled confirmation card for a remote app\npython3 scripts/cli.py prefill \"\u003clink-or-id\u003e\" --prompt \"\u003crequest\u003e\"\n\n# Submit execution for a specific App\npython3 scripts/cli.py run \u003cAPP_ID\u003e --prompt \"\u003cprompt\u003e\"\n```\n\n**ModelZoo model services:**\n```bash\n# Search ModelZoo endpoints\npython3 scripts/cli.py modelzoo-list --keyword \"\u003ckeyword\u003e\"\n\n# View endpoint details and pricing\npython3 scripts/cli.py modelzoo-detail \u003cendpoint\u003e\npython3 scripts/cli.py modelzoo-price \u003cendpoint\u003e\n\n# Execute a ModelZoo task\npython3 scripts/cli.py modelzoo-run \u003cendpoint\u003e --param prompt=\"\u003cprompt\u003e\"\n\n# Query ModelZoo async task status\npython3 scripts/cli.py modelzoo-status \u003crequest_id\u003e\n```\n\n**Account utilities:**\n```bash\n# Validate API key and check balance\npython3 scripts/cli.py check\npython3 scripts/cli.py wallet\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsiliconflow%2Fbizyair-skill","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsiliconflow%2Fbizyair-skill","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsiliconflow%2Fbizyair-skill/lists"}