Kleap
Enables AI agents to build, edit, and publish live websites with hosting, database, auth, and domains via the Model Context Protocol.
README
Kleap — website infrastructure for AI agents
Your agent builds. Kleap ships it live. Let any AI agent — Claude, ChatGPT, Cursor — build, edit and publish real, live websites for you. Hosting, database, auth and domains included.
An alternative to Lovable / v0 / Bolt — except it's driven by your agent, and every publish comes with the verified-live guarantee: a site is only ever reported online once it is provably serving — never a hallucinated dead link.
This is a thin Model Context Protocol server
that wraps Kleap's public REST API. No secrets live in this package — it reads
your own KLEAP_API_KEY from the environment and talks only to kleap.co.
Quick start
1. Get an API key — at kleap.co → Settings → API key →
MCP / API access → Generate MCP key (kleap_live_sk_...).
2. Add Kleap to your AI client:
<details open> <summary><b>Claude Desktop</b> — <code>claude_desktop_config.json</code></summary>
{
"mcpServers": {
"kleap": {
"command": "npx",
"args": ["-y", "github:Kleap-co/kleap"],
"env": { "KLEAP_API_KEY": "kleap_live_sk_..." }
}
}
}
</details>
<details> <summary><b>Cursor</b> — <code>.cursor/mcp.json</code></summary>
{
"mcpServers": {
"kleap": {
"command": "npx",
"args": ["-y", "github:Kleap-co/kleap"],
"env": { "KLEAP_API_KEY": "kleap_live_sk_..." }
}
}
}
</details>
<details> <summary><b>ChatGPT & hosted agents</b> — no local process</summary>
Add the hosted connector at https://kleap.co/api/mcp and authorize with
OAuth (or paste your kleap_live_sk_ key).
</details>
Once published to npm, the install shortens to
"args": ["-y", "kleap"].
3. Restart the client and just ask:
"Build me a one-page site for my bakery, publish it, and give me the live URL." "Add a contact form to my site and redeploy." "Change the headline to 'Roasted slow' and publish."
Works with any MCP-compatible agent: Claude · ChatGPT · Cursor · Claude Code · Codex.
Tools
Build — create_app · modify_app · check_task · retry_task
Publish & domains — publish_app · get_publish_status · search_domains · check_domain · connect_domain
Account — list_apps · get_app · list_app_files · get_credits
| Tool | What it does |
|---|---|
create_app |
Create a site from a prompt → returns a task |
modify_app |
Ask the app's AI to change it → returns a task |
check_task |
Poll a create/modify task to completion |
retry_task |
Resume a failed/stalled build from partial state |
publish_app |
Publish with verified-live (live-or-rollback, never a false "online") |
get_publish_status |
Confirm a site is actually published + live |
search_domains |
Find available domains (purchase stays user-confirmed in Kleap) |
connect_domain |
Connect a domain you already own to a published app |
check_domain |
A domain's connection / DNS status |
list_apps / get_app / list_app_files |
Your apps, an app's details, its files |
get_credits |
Remaining credit balance + plan |
App arguments are snake_case: app_id, task_id, prompt, message, visibility.
The verified-live guarantee
Most tools tell the agent "it's online" the moment a deploy is requested. Kleap reports a site as published only once the new version is provably serving at its live URL — otherwise it rolls back and reports "not confirmed live." Your agent can never hand a user a dead link.
If check_task reports failed (a transient generation stall), call retry_task
with the same task_id to resume from where it stopped — partial work is kept.
FAQ
Is it safe? Yes. The connection uses your private kleap_live_sk_ key; an
agent can only touch your own Kleap apps. The key stays in your local client
config, is sent only to kleap.co over HTTPS, and is revocable anytime in
Settings → API key. Nothing is written to disk.
How much does it cost? Connecting is free. Builds and edits use Kleap credits
(get_credits reports your balance) — see pricing.
Which agents work? Any MCP client: Claude Desktop, Claude Code, Cursor, ChatGPT (hosted connector), and others.
Requirements & run
Node ≥ 18. Run it directly:
KLEAP_API_KEY=kleap_live_sk_... npx -y github:Kleap-co/kleap
# → [kleap-mcp] ready (stdio) → https://kleap.co. Tools: list_apps, ...
Override the API base with KLEAP_API_URL (default https://kleap.co).
Missing key → the server exits with a clear message.
Links
- Kleap: https://kleap.co · MCP & CLI page: https://kleap.co/mcp
- Issues & security: https://github.com/Kleap-co/kleap/issues
Maintained by the Kleap team. MIT © Kleap.
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。