convez.ai
MCP ServerIntegração & APIsIntermediário
HiAPIAI

HiAPI Skills — Diretório de Skills para Agentes IA

HiAPIAI/hiapi-skills

Skills prontos para agentes IA com geração de imagens e vídeos

Ver no GitHub
InicialPluginScore · de 0 a 100

Diretório oficial de skills HiAPI para agentes IA (Claude Code, Cursor, OpenClaw) com suporte a geração de imagens e vídeos via APIs compatíveis com OpenAI. Oferece skills focados em workflows específicos de modelos, integração MCP remota e documentação completa para agents e serviços backend.

Instalação

$npx convez-ai HiAPIAI/hiapi-skills

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
2
Acompanhando
0
Saúde
50%
Último commit
há 1 dia
Versão
—
Issues abertas
3
Contribuidores
1

Casos de uso

  • Integrar geração de imagens e vídeos em agents IA
  • Usar HiAPI com Claude Code, Cursor ou OpenClaw
  • Construir workflows de IA automatizados para criação de conteúdo visual

Veredito

MCP server bem estruturado que conecta agentes IA a capacidades de geração de imagens e vídeos via HiAPI. Recomendado para desenvolvedores que precisam de integração rápida com modelos de IA específicos em workflows de agents.

README

A tradução para o português deste README ainda está na fila.

HiAPI Skills

Official HiAPI skill directory for AI Agents.

HiAPI Skills • Install • API Key • HiAPI

Get API Key · Pricing · Docs · GPT Image 2 Prompts · Seedance 2.0 Prompts · Remote MCP Guide · 简体中文

HiAPI Matrix: 🎨 Image Prompts · 🎬 Video Prompts · 🛠️ Agent Skills (you are here) · 🤖 Remote MCP · 📖 API Docs


AI Agent? Start with llms-install.md. It tells your agent which skill to install, how to set HIAPI_API_KEY, and what to do when a key, balance, quota, or model request fails.


What Is This?

This is the official directory of HiAPI skills for OpenClaw, Claude Code, Codex, OpenCode, Cursor-style agent workflows, and other tools that can read local skills. It helps you choose the right HiAPI entry point before installing a specific skill.

HiAPI is an AI API platform built for developers: one API, all AI models. These skills package common image and video generation workflows so an AI Agent can call a focused model without guessing endpoints or parameters.

Use this repository when you want to choose a HiAPI skill. Install the individual skill repository when you are ready to generate images or videos.


Public Entry Map

EntryLinkUse it when...
Image Prompt Galleryawesome-gpt-image-2-promptsYou want output-backed GPT Image 2 recipes before generating images.
Video Prompt Galleryawesome-seedance-2-0-promptsYou want output-backed Seedance 2.0 recipes for text-to-video or image-to-video.
Agent Skillshiapi-skillsYou want to browse available HiAPI skills and choose one focused workflow.
Remote MCPhttps://mcp.hiapi.ai/mcpYour client supports remote MCP and can pass Authorization: Bearer <HIAPI_API_KEY>.
Async Tasks APIdocs.hiapi.ai/api-referenceYou are building a backend service that needs POST /v1/tasks, polling, signed callbacks, and task history.
API Cookbookdocs.hiapi.aiYou want direct API request shapes, model parameters, and integration guides.

These entry points solve different jobs:

  • Prompt Galleries — output-backed recipes for creators and developers who want a working starting point before calling an API.
  • Agent Skills — installable single-model workflows for agents that need a stable model-specific tool, plus a prompt-only director that turns briefs into runnable video prompts.
  • Remote MCP — a hosted MCP endpoint for clients that can pass a HiAPI API key in request headers.
  • Async Tasks API — the durable /v1/tasks contract for product integrations, queues, polling, signed callbacks, and task history.
  • API Cookbook — docs and examples for direct API integration.

Available Skills

SkillBest ForModelRepository
GPT Image 2Posters, illustrations, social graphics, product visuals, cover imagesgpt-image-2hiapi-gpt-image-2-skill
Seedance 2.0 VideoText-to-video, image-to-video, cinematic clips, product videos, storyboardsseedance-2-0hiapi-seedance-2-0-video-skill
HappyHorse 1.0 VideoLightweight text-to-video drafts, short social clips, ad conceptshappyhorse-1-0hiapi-happyhorse-1-0-video-skill
Video Prompt GeneratorDirect briefs, links, and topics into scene-by-scene Seedance/HappyHorse prompts before generating— (prompt-only)hiapi-video-prompt-generator-skill

Choose an Entry

I want to...Open
Find a tested image prompt and output exampleAwesome GPT Image 2 Prompts
Find a tested video prompt and output exampleAwesome Seedance 2.0 Prompts
Generate images from textGPT Image 2 Skill
Generate or animate videos with a stronger video workflowSeedance 2.0 Video Skill
Quickly generate short text-to-video clipsHappyHorse 1.0 Video Skill
Turn a one-line brief or link into a scene-by-scene video prompt before generatingVideo Prompt Generator Skill
Let an agent access more HiAPI models from chatHiAPI Remote MCP Guide
Build a server-side integration with webhooks and task historyAsync Tasks API

Skills are best when you want a stable, focused workflow. Remote MCP is better when you want a chat agent to discover and call multiple HiAPI tools. The MCP endpoint is https://mcp.hiapi.ai/mcp.

For application code, use the direct /v1/tasks API: submit with POST https://api.hiapi.ai/v1/tasks, poll GET https://api.hiapi.ai/v1/tasks/{taskId}, or register callback.url to receive terminal-state webhooks signed by X-HiAPI-Timestamp and X-HiAPI-Signature.


MCP vs Skill

Both routes use the same HiAPI account and the same HIAPI_API_KEY. Pick by how your agent runs and how stable you want the workflow.

DimensionSingle-Model SkillRemote MCP
What it isA local skill folder loaded by your agent (Codex, Claude Code, OpenClaw, …)A hosted MCP endpoint at https://mcp.hiapi.ai/mcp
Install pathgit clone or openclaw skills add into the agent's skills directoryAdd one JSON block to the agent's MCP config — no clone required
NetworkDirect HTTPS to api.hiapi.ai from the user's machineAgent → mcp.hiapi.ai (hosted MCP) → api.hiapi.ai
Best forStable, focused, single-model workflows — image-only or video-onlyChat agents that need to discover and call multiple HiAPI tools in one session
Model surfaceOne model per skill (gpt-image-2, seedance-2-0, happyhorse-1-0)All MCP-exposed tools: generate_image, generate_video, list_models, …
UpdatesYou pull from the repo when there's a new versionHosted — capabilities update server-side
Required client supportAny agent that can read a local skills folderClient must support remote MCP URLs and custom headers (Authorization: Bearer …)
Image uploadLocal file paths supported by the skill scriptsUpload via URL — local-file upload is being expanded
Use it whenThe model and use case are fixed and you want zero configuration driftYou don't know in advance which tool the agent will need

If both fit, install a single-model skill first — it has the smallest moving parts. Add Remote MCP later when you need a wider tool surface in chat.

Skills vs Remote MCP vs /v1/tasks

RouteChoose it whenPrimary interface
SkillsThe user asked for one known model or a repeatable local workflow.Local skill scripts plus HIAPI_API_KEY
Remote MCPThe user wants an agent to discover models and tools in chat.https://mcp.hiapi.ai/mcp
Direct /v1/tasksYou are building product code, queues, automations, or server-side callbacks.POST /v1/tasks, GET /v1/tasks/{taskId}, GET /v1/tasks

The async task lifecycle is queued -> handling -> archiving -> success or fail. Successful tasks return output[] assets with url, type, and expireAt. Failed tasks return error.code, including INVALID_REQUEST, MODEL_UNAVAILABLE, TASK_FAILED, TASK_TIMEOUT, and STORAGE_UNAVAILABLE.


Install One Skill

Pick the repository for the model you need. Install all three only if you want all three skills available in your agent.

One Command (Recommended)

# GPT Image 2
npx -y github:HiAPIAI/hiapi-gpt-image-2-skill -y

# Seedance 2.0 Video
npx -y github:HiAPIAI/hiapi-seedance-2-0-video-skill -y

# HappyHorse 1.0 Video
npx -y github:HiAPIAI/hiapi-happyhorse-1-0-video-skill -y

# Video Prompt Generator (prompt-only, no API key needed)
npx -y github:HiAPIAI/hiapi-video-prompt-generator-skill -y

Each installer auto-detects Codex (~/.codex/skills) and Claude Code (~/.claude/skills). Pass --codex, --claude, --target=/path, or set AGENT_SKILLS_DIR to override. Requires Node 18+ and git on PATH.

OpenClaw

# GPT Image 2
openclaw skills add https://github.com/HiAPIAI/hiapi-gpt-image-2-skill

# Seedance 2.0 Video
openclaw skills add https://github.com/HiAPIAI/hiapi-seedance-2-0-video-skill

# HappyHorse 1.0 Video
openclaw skills add https://github.com/HiAPIAI/hiapi-happyhorse-1-0-video-skill

# Video Prompt Generator
openclaw skills add https://github.com/HiAPIAI/hiapi-video-prompt-generator-skill

Codex

mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills"

rm -rf "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-gpt-image-2"
git clone https://github.com/HiAPIAI/hiapi-gpt-image-2-skill.git "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-gpt-image-2"

rm -rf "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-seedance-2-0-video"
git clone https://github.com/HiAPIAI/hiapi-seedance-2-0-video-skill.git "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-seedance-2-0-video"

rm -rf "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-happyhorse-1-0-video"
git clone https://github.com/HiAPIAI/hiapi-happyhorse-1-0-video-skill.git "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-happyhorse-1-0-video"

rm -rf "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-video-prompt-generator"
git clone https://github.com/HiAPIAI/hiapi-video-prompt-generator-skill.git "${CODEX_HOME:-$HOME/.codex}/skills/hiapi-video-prompt-generator"

Restart Codex after installing skills.

Claude Code

mkdir -p "$HOME/.claude/skills"

rm -rf "$HOME/.claude/skills/hiapi-gpt-image-2"
git clone https://github.com/HiAPIAI/hiapi-gpt-image-2-skill.git "$HOME/.claude/skills/hiapi-gpt-image-2"

rm -rf "$HOME/.claude/skills/hiapi-seedance-2-0-video"
git clone https://github.com/HiAPIAI/hiapi-seedance-2-0-video-skill.git "$HOME/.claude/skills/hiapi-seedance-2-0-video"

rm -rf "$HOME/.claude/skills/hiapi-happyhorse-1-0-video"
git clone https://github.com/HiAPIAI/hiapi-happyhorse-1-0-video-skill.git "$HOME/.claude/skills/hiapi-happyhorse-1-0-video"

rm -rf "$HOME/.claude/skills/hiapi-video-prompt-generator"
git clone https://github.com/HiAPIAI/hiapi-video-prompt-generator-skill.git "$HOME/.claude/skills/hiapi-video-prompt-generator"

Any Agent With A Skills Folder

export AGENT_SKILLS_DIR="/path/to/your/agent/skills"
mkdir -p "$AGENT_SKILLS_DIR"

rm -rf "$AGENT_SKILLS_DIR/hiapi-gpt-image-2"
git clone https://github.com/HiAPIAI/hiapi-gpt-image-2-skill.git "$AGENT_SKILLS_DIR/hiapi-gpt-image-2"

rm -rf "$AGENT_SKILLS_DIR/hiapi-seedance-2-0-video"
git clone https://github.com/HiAPIAI/hiapi-seedance-2-0-video-skill.git "$AGENT_SKILLS_DIR/hiapi-seedance-2-0-video"

rm -rf "$AGENT_SKILLS_DIR/hiapi-happyhorse-1-0-video"
git clone https://github.com/HiAPIAI/hiapi-happyhorse-1-0-video-skill.git "$AGENT_SKILLS_DIR/hiapi-happyhorse-1-0-video"

rm -rf "$AGENT_SKILLS_DIR/hiapi-video-prompt-generator"
git clone https://github.com/HiAPIAI/hiapi-video-prompt-generator-skill.git "$AGENT_SKILLS_DIR/hiapi-video-prompt-generator"

Get API Key

  1. Open HiAPI API Keys.
  2. Create or sign in to your HiAPI account.
  3. Create or copy an API key.
  4. Set it in the terminal where your agent runs:
export HIAPI_API_KEY="your_hiapi_api_key_here"
export HIAPI_BASE_URL="https://api.hiapi.ai"

Each skill includes a scripts/check-config.mjs helper so your agent can verify the key and network path before generation.


Agent Prompt

Copy this to your AI Agent:

Use HiAPI skills for image and video generation.

If I ask for an image, install or use:
https://github.com/HiAPIAI/hiapi-gpt-image-2-skill

If I ask for a stronger video workflow or image-to-video, install or use:
https://github.com/HiAPIAI/hiapi-seedance-2-0-video-skill

If I ask for a quick text-to-video clip, install or use:
https://github.com/HiAPIAI/hiapi-happyhorse-1-0-video-skill

Use HIAPI_API_KEY for authentication. If the key is missing, invalid, rate-limited, or the account has insufficient balance, show the user the next step and link to HiAPI.

Prompt Gallery Workflow

Use awesome-gpt-image-2-prompts when the user needs inspiration, style references, or a prompt starting point. Each case includes a result image, source attribution, aspect ratio, and a HiAPI Draw link. After the user chooses a case, use the relevant skill or direct API request to generate a new result with their own subject, product, text, or scene.


Machine-Readable Index

This repository includes skills.json, a machine-readable index of the current HiAPI skill repositories, model pages, install commands, and intended use cases.


Troubleshooting

ProblemNext Step
Missing HIAPI_API_KEYCreate or copy a key at HiAPI API Keys, then export HIAPI_API_KEY.
Invalid key or unauthorized requestCheck the key or create a new one at HiAPI API Keys.
Insufficient balance, quota, or payment statusCheck your account in the HiAPI Dashboard and review Pricing.
Rate limitedWait and retry, or reduce concurrent generations.
Unsure which model to useStart with the table above, or follow the HiAPI Remote MCP Guide so your agent can inspect available tools.

Related HiAPI Repositories


License

MIT


HiAPI — One API, all AI models.

Ver o original no GitHub

Bom para

Relacionados