NormaHub MCP Server
OfficialNormaHub MCP Server
Сервер Model Context Protocol (MCP) для API NormaHub.
Предоставляет инструменты (Tools), ресурсы (Resources) и шаблоны промптов (Prompts) для подключения моделей (DeepSeek, Claude, GPT, SeeDance Video) к Claude Desktop, Cursor, Windsurf и другим MCP-клиентам.
Поддерживает два вида транспорта:
Stdio — локальный запуск через
npxв IDE и десктопных клиентах.SSE (Server-Sent Events) — сетевой HTTP-транспорт для Docker, Kubernetes и удаленных серверов.
Архитектура сервера
┌────────────────────────────────────────────────────────┐
│ NormaHub MCP Server │
└───────────────────────────┬────────────────────────────┘
│
┌──────────────────────┼──────────────────────┐
▼ ▼ ▼
TOOLS RESOURCES PROMPTS
• normahub_prompt_model• normahub://models • review_code
• normahub_image_to_video • normahub://status • direct_video_scene
• normahub_generate_video • cost_optimize_route
• normahub_generate_image
• normahub_list_modelsRelated MCP server: ComfyUI MCP Server
Возможности
1. Tools (Инструменты)
Инструмент | Описание | Основные параметры |
| Вызов LLM (диалоги и история сообщений) |
|
| Генерация видео из статичного изображения (I2V) через SeeDance |
|
| Генерация видео из текста (T2V) через SeeDance / Grok Video |
|
| Генерация изображений |
|
| Список подключенных моделей и тарифов | — |
2. Resources (Ресурсы)
normahub://models— структурированный каталог поддерживаемых моделей, контекстных окон, capabilities и тарифов (JSON и Markdown).normahub://status— статус шлюза, latency ping и список активных маршрутов.
3. Prompts (Шаблоны)
review_code— шаблон для код-ревью (архитектура, безопасность, производительность, diff).direct_video_scene— шаблон для формирования сценария генерации видео (оптика, свет, физика камеры) под SeeDance.cost_optimize_route— подбор наиболее экономичной модели под задачу.
Быстрый старт
Claude Desktop (Stdio)
Добавьте сервер в конфигурационный файл Claude Desktop:
macOS:
~/Library/Application Support/Claude/claude_desktop_config.jsonLinux:
~/.config/Claude/claude_desktop_config.jsonWindows:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"normahub": {
"command": "npx",
"args": ["-y", "@normahub/mcp-server"],
"env": {
"NORMAHUB_API_KEY": "nr_hub_..."
}
}
}
}Cursor IDE
В настройках Settings > Features > MCP > Add New MCP Server:
Name:
normahubType:
commandCommand:
npx -y @normahub/mcp-serverEnvironment Variables:
NORMAHUB_API_KEY=nr_hub_...
Готовые шаблоны конфигураций
Готовые JSON-файлы конфигурации для Claude Desktop, Cursor (.cursor/mcp.json), Zed и Continue.dev доступны в директории examples/.
Сетевой запуск (SSE Transport)
Для совместного использования командой или работы в контейнерах:
Локальный запуск SSE:
git clone https://github.com/normahubcc/mcp-server-normahub.git
cd mcp-server-normahub
npm install
npm run build
export NORMAHUB_API_KEY="nr_hub_..."
npm run start:sseСервер будет доступен на порту 3001:
SSE поток:
http://localhost:3001/ssePOST сообщений:
http://localhost:3001/messagesHealth check:
http://localhost:3001/health
Запуск в Docker:
docker build -t normahub-mcp-server .
docker run -d \
-p 3001:3001 \
-e NORMAHUB_API_KEY="nr_hub_..." \
--name normahub-mcp \
normahub-mcp-serverОтладка через MCP Inspector
npm run inspectКоманда запустит локальный веб-интерфейс MCP Inspector для интерактивного тестирования вызовов инструментов и чтения ресурсов.
Лицензия
MIT License. См. LICENSE.
Available Tools
7 toolsnormahub_generate_imageC
Generate images using GPT Image or Grok Imagine Image via NormaHub API.
Inputs: prompt (required), model.
normahub_generate_videoB
Generate AI videos using SeeDance 2.5 / 2.0 and xAI Grok Imagine Video via NormaHub API. Automatically polls until video render completes.
Inputs: prompt (required), model.
normahub_get_balanceA
Check your current NormaHub account balance in Russian Rubles (RUB), active tier, and remaining tokens.
Inputs: no named parameters.
normahub_image_to_videoB
Animate an input image into a video (Image-to-Video) using SeeDance 2.5 / 2.0 via NormaHub API. Accepts image URL and animation prompt. Automatically polls until video render completes.
Inputs: prompt (required), image_url (required), model.
normahub_list_modelsA
List currently available active AI models and their pricing on NormaHub API.
Inputs: no named parameters.
normahub_prompt_modelC
Send a prompt to LLM models via NormaHub API (Claude, GPT, DeepSeek, Qwen).
Inputs: model, prompt, system, messages, +2 more.
normahub_transcribe_audioA
Transcribe audio recordings, voice notes, and video soundtracks to text using OpenAI Whisper via NormaHub API without VPN. Accepts local file path or public HTTP URL.
Inputs: url, model, prompt, language, +1 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 7 tools
Each tool maps to a distinct capability: text prompting, image generation, text-to-video, image-to-video, transcription, model listing, and billing. The only mild overlap is between normahub_generate_video and normahub_image_to_video, but the descriptions clearly differentiate them by input (image required vs. prompt-only).
Every tool uses the same normahub_ prefix followed by a clear verb_noun or verb_object pattern in snake_case. No mixing of conventions or vague standalone verbs.
Seven tools is a well-scoped surface for a multi-modal AI API gateway. Each tool corresponds to a real, non-redundant operation and no filler tools are present.
The surface covers the core modalities (text, image, video, image-to-video, audio transcription) plus account inspection via balance and model listing. Minor gaps exist: no text-to-speech/audio generation, no image editing/upscale, and no usage-history or job-status tool, though built-in polling mitigates the latter.
Maintenance
Related MCP Connectors
Image and video AI tools and your own pipelines, run from any AI assistant.
Generate AI images, video, voiceovers and music from Claude, ChatGPT or Cursor through 50+ models (Veo 3.1, Kling 3, Seedance, Nano Banana, GPT Image, ElevenLabs). Also image editing, upscaling, background removal, face swap, transcription, voice cloning and UGC-style video ads. Sign in with OAuth — no API key to paste. Tools are annotated (read-only vs. credit-spending); failed generations are refunded.
Generate and edit images, video, voice, lip-sync and 3D models from your AI agent.
One workspace of tools for Claude and ChatGPT: connect 600+ apps, generate media, build tools.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI agents to connect to and chat with multiple LLM models (OpenAI, OpenRouter, custom endpoints) with conversation history management and model switching capabilities.318 npm2MIT
- AlicenseNot gradedqualityNot gradedmaintenanceEnables AI agents to generate and iteratively refine images, audio, and video by interacting with a local ComfyUI instance through natural conversation. It provides comprehensive tools for workflow management, node introspection, and publishing generated assets.Apache 2.0
- FlicenseBqualityDmaintenanceConnects AI assistants to ComfyUI for image, video, and audio generation, providing full control over ComfyUI through 40+ tools including quick generation, cloud API nodes, and custom workflow building.46-
- AlicenseAqualityCmaintenanceEnables agents to generate images, video, audio, and documents via AihubMax API, with task management and media analysis capabilities.14MIT