jdownloader-mcp
MCP server to drive a local JDownloader 2 instance through its entire download lifecycle, including adding links/containers, linkgrabber management, queue control, priority, force-start, progress tracking, and cleanup.
README
jdownloader-mcp
An MCP server that lets an agent drive a local JDownloader 2 through its entire download lifecycle — add links / containers, watch the linkgrabber resolve them, move them to the queue, set priority, force-start, track progress, inspect download/extract directories, and clean up.
It exists for the one download workload the shell can't already do: premium
hosters, one-click hosters, .dlc/.crypt containers, and captcha-gated folder
links. (Plain URLs and video are better served by aria2c / yt-dlp.)
Thin wrapper over myjdapi;
built on the mcp 1.x FastMCP API.
Status: v0.1.0 — untested against a live JDownloader. Passes 52 mocked unit tests and a real MCP stdio handshake, but the end-to-end lifecycle (
smoke_test.py) has not yet been run against a live JD2. Treat as a pre-release; verify before relying on it.
Lifecycle
add_links / add_container ─▶ LINKGRABBER ─(move_to_downloads)─▶ DOWNLOADS ─▶ extract ─▶ remove
list_linkgrabber list_downloads / force_start / wait_for
Item IDs (uuid) come from the list_* tools; most tools take
package_ids / link_ids plus a stage (linkgrabber | downloads).
Tools
add_links · add_container · list_linkgrabber · list_downloads ·
status · wait_for · move_to_downloads · control_downloads ·
force_start · set_priority · set_enabled · get_directories ·
set_download_dir · set_extraction · remove · reset · raw_action
(generic passthrough to any JD API endpoint not wrapped above).
Setup
- In JDownloader: Settings → My.JDownloader, sign in / link the device, note its device name.
- Configure creds:
cp .env.example .env # then edit .env - Install deps (already done if you used
uv):uv sync
Verify
uv run pytest # mocked unit tests (no JD needed)
uv run python smoke_test.py # live end-to-end against your JD2 (uses .env)
The smoke test walks the whole lifecycle with a tiny file and cleans up. It SKIPs cleanly if the env vars aren't set.
Register with Claude Code
claude mcp add jdownloader -- uv run --directory E:/SWE-Pioneers/projects/jdownloader-mcp jdownloader-mcp
or add to your MCP client config (stdio):
{
"mcpServers": {
"jdownloader": {
"command": "uv",
"args": ["run", "--directory", "E:/SWE-Pioneers/projects/jdownloader-mcp", "jdownloader-mcp"]
}
}
}
The server reads MYJD_EMAIL / MYJD_PASSWORD / MYJD_DEVICE from .env (or
the environment).
Docs
docs/2026-08-01-jdownloader-mcp-design.md— design spec.docs/myjdapi-api-reference.md— exact installed-API reference the code targets.
License
MIT — see LICENSE.
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。