Dribba

Dribba

Official SDK, CLI, MCP server and agent skills for the Dribba public API — no API key required.

Category
访问服务器

README

dribba

Official SDK and CLI for the Dribba public API.

No API key, no OAuth, no account. Every read endpoint is public. If you were looking for the credentials page, there isn't one — see dribba.com/auth.md.

CLI

Nothing to install:

npx dribba services --table
npx dribba cases --limit 3 | jq -r '.items[].slug'
npx dribba estimate --platforms ios,android --complexity complex
npx dribba ask "flutter migration"
npx dribba markdown /servicios
npx dribba --help

Add --sandbox to hit the frozen fixtures instead of production.

MCP server (stdio bridge)

Dribba's MCP server is remote — Streamable HTTP at https://dribba.com/mcp. If your client only speaks stdio, this package ships the bridge:

{
  "mcpServers": {
    "dribba": { "command": "npx", "args": ["-y", "dribba-mcp"] }
  }
}
npx dribba-mcp            # product surface: 5 tools + llms.txt resources
npx dribba-mcp --docs     # documentation surface: 4 tools

It forwards JSON-RPC and nothing else, so the tools, resources and protocol version are whatever the server declares — there is no second implementation to drift. If your client does speak Streamable HTTP, skip the bridge and point it straight at the URL.

Verified with the official @modelcontextprotocol/sdk over stdio: connect, tools/list, tools/call, resources/list, resources/read.

SDK

npm i dribba        # JavaScript / TypeScript
pip install dribba  # Python 3.9+, same methods, also dependency-free

The Python client lives in python/ and is published to PyPI as dribba.

import { Dribba } from "dribba";

const dribba = new Dribba();

const { items, total, next_cursor } = await dribba.services({ limit: 5 });
const study = await dribba.case("cityxerpa");
const budget = await dribba.estimate({ platforms: ["ios", "android"] });

// Whole collections, following the cursor for you
for await (const job of dribba.paginate("/api/v1/jobs")) {
  console.log(job.title);
}

Errors

Every 4xx/5xx throws a DribbaError. Branch on code — it is stable. title and detail are prose and may be reworded.

import { DribbaError } from "dribba";

try {
  await dribba.service("nope");
} catch (error) {
  if (error instanceof DribbaError && error.code === "resource_not_found") {
    console.log(error.resolution); // what to do next, from the server
  }
}

Rate limits

120 requests per 60 seconds per IP. After any call, dribba.rateLimit holds what the last response advertised, so you can self-throttle instead of provoking a 429:

await dribba.company();
console.log(dribba.rateLimit); // { limit: 120, remaining: 118, reset: 47 }

Sandbox

Frozen fixtures with production shapes. Useful because a test that asserts "there are 9 services" breaks the day we publish the tenth — against the sandbox it does not.

const sandbox = new Dribba({ sandbox: true });
const services = await sandbox.services(); // services.sandbox === true

Idempotency

Idempotency-Key on any POST. Same key and same body replays the stored response; same key with a different body is a 400.

await dribba.estimate(input, { idempotencyKey: crypto.randomUUID() });

Markdown, and the async export

Any page as markdown, and a one-request export of many pages:

const page = await dribba.markdown("/servicios");

const job = await dribba.startExport(["/", "/precios", "/servicios"]);
await dribba.waitForExport(job.id);
const everything = await dribba.exportResult(job.id);

Export jobs live in one server instance's memory and expire after 15 minutes; a 404 while polling means the request reached a different instance, not that the work was lost.

Beyond the REST API

  • MCP — two remote servers over Streamable HTTP, no auth: https://dribba.com/mcp (actions) and https://dribba.com/docs/mcp (docs). See https://dribba.com/developers/mcp.
  • Markdown of any page — Accept: text/markdown, or append .md.
  • The whole site as an OKF bundle.
  • /ask — NLWeb: a natural-language question, verbatim passages back.

Skills and the Agent Plugins manifest

This repo doubles as a portable Agent Plugin:

plugin.json          the manifest
mcp.json             the two remote MCP servers
skills/              seven SKILL.md files describing what dribba.com exposes

Install the skills into your agent:

npx skills add dribbaengineering/dribba

skills/ is generated from what the site serves at https://dribba.com/.well-known/agent-skills/. Edit the site, not the copy — the copy is what drifts.

Requirements

Node 20+. Zero dependencies.

License

MIT. The data behind the API is CC BY 4.0 — cite "Dribba" and link the source URL.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选