Intelliverse Router MCP Server
Enables agents to manage the Intelliverse platform: create apps, API keys, knowledge bases, send email, generate media, and chat without specifying an LLM.
README
Intelliverse Router MCP server + Agent Skills
Run the whole Intelliverse platform from any MCP-capable agent: create App IDs, mint tier API keys, ingest and search per-app knowledge bases, send email, generate media, check credits, and chat — without ever naming an LLM.
npx -y github:intelli-verse-x/router-mcp
Verified 2026-07-12: initialize + tools/list over stdio, 15 tools.
Full setup guide with copy-paste configs: https://router.intelli-verse-x.ai/skills
Tools (15)
| Tool | Auth | What it does |
|---|---|---|
chat |
API key | Chat completion; model optional (Intelliverse default routes at your plan tier) |
list_models |
API key | Tier aliases your key can call |
get_credits |
API key | LLM daily budget + per-App-ID media credit wallets |
create_app / list_apps |
management | App IDs — entry point for KB, wallets, per-app costs |
create_api_key / list_api_keys / revoke_api_key |
management | Tier-gated key administration |
kb_ingest / kb_search / kb_chat / kb_documents |
API key | App-scoped pgvector knowledge base |
send_email |
API key | Per-App-ID email via the platform engine (Pro+, email.send10 SKU, hold → settle) |
generate_media / media_job_status |
API key | Credit-metered image/video/voice/audio/3D jobs |
Env vars: ROUTER_API_KEY (Workspace → API Keys), ROUTER_MANAGEMENT_KEY
(Account → Management Keys; needed for app/key admin tools),
ROUTER_WORKSPACE_ID (default workspace for admin tools). Optional overrides:
EDGE_BASE_URL, KB_BASE_URL. Keys are free —
sign up, no card.
Install per platform
The JSON body is identical everywhere; only the file location / wrapper differs.
Cursor
One-click: Add to Cursor
— or .cursor/mcp.json (project) / ~/.cursor/mcp.json (global):
{
"mcpServers": {
"intelliverse-router": {
"command": "npx",
"args": ["-y", "github:intelli-verse-x/router-mcp"],
"env": {
"ROUTER_API_KEY": "iv-r-v1-...",
"ROUTER_MANAGEMENT_KEY": "ivm-...",
"ROUTER_WORKSPACE_ID": "your-workspace-uuid"
}
}
}
}
Claude Desktop
Same JSON as Cursor, in claude_desktop_config.json
(macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\).
Claude Code
claude mcp add intelliverse-router \
--env ROUTER_API_KEY=iv-r-v1-... \
--env ROUTER_MANAGEMENT_KEY=ivm-... \
--env ROUTER_WORKSPACE_ID=your-workspace-uuid \
-- npx -y github:intelli-verse-x/router-mcp
VS Code (GitHub Copilot)
.vscode/mcp.json:
{
"servers": {
"intelliverse-router": {
"type": "stdio",
"command": "npx",
"args": ["-y", "github:intelli-verse-x/router-mcp"],
"env": {
"ROUTER_API_KEY": "iv-r-v1-...",
"ROUTER_MANAGEMENT_KEY": "ivm-...",
"ROUTER_WORKSPACE_ID": "your-workspace-uuid"
}
}
}
}
Windsurf
Same JSON as Cursor, in ~/.codeium/windsurf/mcp_config.json.
Cline
Same JSON as Cursor, via the MCP Servers panel → Configure MCP Servers
(cline_mcp_settings.json).
Continue
~/.continue/config.yaml:
mcpServers:
- name: intelliverse-router
command: npx
args: ["-y", "github:intelli-verse-x/router-mcp"]
env:
ROUTER_API_KEY: iv-r-v1-...
ROUTER_MANAGEMENT_KEY: ivm-...
ROUTER_WORKSPACE_ID: your-workspace-uuid
OpenAI Codex CLI
codex mcp add intelliverse-router \
--env ROUTER_API_KEY=iv-r-v1-... \
--env ROUTER_MANAGEMENT_KEY=ivm-... \
--env ROUTER_WORKSPACE_ID=your-workspace-uuid \
-- npx -y github:intelli-verse-x/router-mcp
ChatGPT
ChatGPT developer-mode apps connect only to remote MCP servers over HTTPS; this server is stdio-only today, so there is no ChatGPT path yet (hosted streamable-HTTP endpoint is on the roadmap). In the OpenAI ecosystem use the Codex CLI above.
Agent Skills
skills/ ships SKILL.md playbooks that drive these tools end-to-end — copy them
into your agent's skills folder (~/.cursor/skills/ for Cursor,
~/.claude/skills/ for Claude Code):
| Skill | What it does |
|---|---|
intelliverse-router |
Platform wiring — the 60-second create_app → key → KB loop |
nl-app-builder |
Scaffold web/mobile apps wired to an App ID |
scaffold-game |
Honest-scope Nakama game-backend stubs |
deploy-desktop |
Electron/Tauri desktop apps delegating to IVX Agency |
conversational-avatar |
Real-time voice + lip-sync avatars over LiveKit |
git clone https://github.com/intelli-verse-x/router-mcp
cp -r router-mcp/skills/* ~/.cursor/skills/ # or ~/.claude/skills/
The 60-second loop
create_app— describe your app; get an App ID and recommended tier.create_api_key— mint a key bound to that app.kb_ingest— feed it docs or URLs.chat/kb_chat— start building. No model names, ever.send_email— tell your users it's live.
Development
npm install && npm run build && node dist/index.js
server.json at the repo root is the payload published to the official
MCP Registry as
io.github.intelli-verse-x/router-mcp (MCPB bundle from the GitHub release).
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。