melopulse
MeloPulse is an offline-first playlist recommender for coding sessions, exposing MCP tools to recommend playlists, add playlist links, list local catalogue entries, and sync the MeloLab public catalogue.
README
MeloPulse
Local coding playlists for MCP agents.
npx @melolab/melopulse recommend
MeloPulse reads safe Git metadata locally, recommends from a local catalogue, and opens playlists in MeloLab or your music app. Recommendation does not upload code or require an account.
What it does
MeloPulse is an offline-first playlist recommender for coding sessions. It includes a small local catalogue, lets you import HTTPS playlist links with local tags, and can explicitly synchronize MeloLab's public catalogue into a local cache. It returns playlist URLs; your browser or music app handles playback.
It does not generate music, store audio, authenticate users, use OAuth, control a player, monitor repositories in the background, send telemetry, or fetch provider metadata. The only network operation is an explicit sync request to MeloLab.
Install and use
The package is not published until a separate release action occurs. Until then, run from a local checkout with npm run build and node dist/cli/index.js <command>.
Once published, use the CLI with npx:
npx -y @melolab/melopulse recommend --activity debugging
npx -y @melolab/melopulse recommend --no-git --json
CLI commands
| Command | Purpose |
|---|---|
melopulse add <playlist-url> |
Import an HTTPS playlist link into the local catalogue. Use --title, --activity, --mood, --energy, --focus, and --vocals to set local tags. |
melopulse sync |
Explicitly download MeloLab's public featured catalogue and update the local cache. |
melopulse recommend |
Return up to three local recommendations. Git context is enabled by default; use --no-git to disable it. |
melopulse play <playlist-id> |
Hand a selected playlist URL to the operating system's default browser or registered music app. |
melopulse mcp |
Start the MCP server over standard input/output. |
recommend accepts --activity, --mood, --energy, --focus, --vocals, --limit, and --json. Recommendation uses the bundled catalogue, your local link imports, and any cached MeloLab sync; it does not query streaming providers.
MCP tools
The MCP server exposes exactly four tools:
| Tool | Purpose |
|---|---|
melopulse_recommend |
Recommend from the local catalogue, optionally using safe Git context. |
melopulse_add_playlist |
Import and tag an HTTPS playlist link locally. |
melopulse_list_playlists |
List local catalogue entries, optionally filtered by source. |
melopulse_sync_catalog |
Explicitly synchronize MeloLab's public catalogue. |
See MCP setup and copy the ready-made configurations in examples.
Supported playlist links
MeloPulse accepts HTTPS links and recognizes MeloLab, Spotify, Apple Music, and YouTube Music links. Other HTTPS playlist links are stored as generic. Link import only saves the URL and your local tags; it does not retrieve provider metadata or audio.
Local storage and privacy
The local catalogue is stored in the platform configuration directory, or in MELOPULSE_CONFIG_DIR when set. It contains playlists.json for local imports and melolab-catalog-cache.json for the last explicit MeloLab sync.
When Git context is enabled, MeloPulse reads only safe metadata from the chosen workspace. It never uploads repository contents, source code, file contents, remote URLs, credentials, or audio. Read the complete privacy boundary.
MeloLab relationship
MeloPulse is a local recommender that can hand off a MeloLab playlist URL to your browser or app. MeloLab sync is opt-in and happens only when you run melopulse sync or call melopulse_sync_catalog; there is no background synchronization, account connection, OAuth, or provider integration.
Development
npm ci
npm run verify
npm run smoke:pack
npm pack --dry-run
The packed-install smoke test builds an npm tarball, installs it in a temporary directory, runs recommend --no-git --json, verifies a recommendation is returned, and cleans up both locations.
Contributing and security
Please read CONTRIBUTING.md, CODE_OF_CONDUCT.md, and SECURITY.md before participating.
License
MIT © 2026 Tover0314
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。