music163-export

# NetEase Cloud Music Playlist Export Export all playlists and songs from a logged-in NetEase Cloud Music account to a structured text file. ## Prerequisites - User must be logged in at music.163.com (see `music163-playlist` skill for auth strategies) - `chrome-devtools-mcp` must be available - Page must have `window.asrsea` loaded (navigate to music.163.com first) ## Strategy: Browser-Side Aggregation **Problem:** `evaluate_script` has a return value size limit (~50KB). Exporting 37 playlists / 4000 songs produces ~100KB of text. Naive approach (return all data in one call) fails. **Solution:** Run the entire export pipeline inside the browser in phases: 1. **Phase 1 — Collect** (single evaluate_script): Fetch all playlist metadata + songs, store in `window.__exportData` 2. **Phase 2 — Format** (single evaluate_script): Generate the text content, store in `window.__musicTxt` 3. **Phase 3 — Transfer** (chunked reads): Extract text in <=40KB chunks via multiple evaluate_script calls 4. **Phase 4 — Write** (local): Concatenate chunks and write to disk ### Phase 1: Collect All Data Execute this in `evaluate_script`. It handles: - Fetching all playlists for the user - Fetching tracks for each playlist (with 300ms rate limiting) - Pagination for large playlists (>1000 tracks) via trackIds + song/detail fallback ```javascript (async () => { const csrf = (document.cookie.match(/__csrf=([^;]+)/) || [])[1] || ""; if (!csrf) return { error: "Not logged in" }; if (typeof window.asrsea !== 'function') return { error: "asrsea not loaded" }; const PK = "010001"; const MOD = "00e0b509f6259df8642dbc35662901477df22677ec152b5ff68ace615bb7b725152b3ab17a876aea8a5aa76d2e417629ec4ee341f56135fccf695280104e0312ecbda92557c93870114af6c9d05c4f7f0c3685b7a46bee255932575cce10b424d813cfe4875d3e82047b97ddef52741d546b8e289dc6935b3ece0462db0a22b8e7"; const NONCE = "0CoJUm6Qyw8W8jud"; async function post(path, data) { data.csrf_token = csrf; const enc = window.asrsea(JSON.stringify(data), PK, MOD, NONCE); const r = await fetch(`https://music.163.com/weapi${path}?csrf_token=${csrf}`, { method: "POST", headers: {"Content-Type": "application/x-www-form-urlencoded"}, credentials: "include", body: `params=${encodeURIComponent(enc.encText)}&encSecKey=${enc.encSecKey}` }); return r.json(); } const delay = ms => new Promise(r => setTimeout(r, ms)); // 1. Get uid const acctResp = await fetch("https://music.163.com/api/nuser/account/get", { method: "POST", credentials: "include", headers: {"Content-Type": "application/x-www-form-urlencoded"}, body: "csrf_token=" + csrf }); const acct = await acctResp.json(); const uid = acct.account?.id; const nickname = acct.profile?.nickname || "Unknown"; if (!uid) return { error: "Cannot get uid" }; // 2. Get all playlists const plResp = await fetch(`https://music.163.com/api/user/playlist/?uid=${uid}&offset=0&limit=1000`, { credentials: "include" }); const plData = await plResp.json(); const playlists = plData.playlist || []; // 3. Fetch tracks for each playlist window.__exportData = {}; window.__exportMeta = { uid, nickname }; for (const pl of playlists) { const detail = await post("/v6/playlist/detail", { id: pl.id, n: 2000 }); if (detail.code !== 200) { await delay(500); continue; } let tracks = (detail.playlist?.tracks || []).map(t => ({ id: t.id, name: t.name, artist: (t.ar || []).map(a => a.name).join(" / ") })); // Handle pagination: if trackIds has more than tracks returned const allIds = (detail.playlist?.trackIds || []).map(t => t.id); if (tracks.length < allIds.length) { const missing = allIds.slice(tracks.length); for (let i = 0; i < missing.length; i += 500) { const batch = missing.slice(i, i + 500); const res = await post("/v3/song/detail", { c: JSON.stringify(batch.map(id => ({id}))), ids: JSON.stringify(batch) }); if (res.songs) { tracks.push(...res.songs.map(t => ({ id: t.id, name: t.name, artist: (t.ar || []).map(a => a.name).join(" / ") }))); } await delay(300); } } window.__exportData[pl.id] = { name: pl.name, tracks, specialType: pl.specialType }; await delay(300); } const totalTracks = Object.values(window.__exportData).reduce((s, p) => s + p.tracks.length, 0); return { status: "ok", playlists: playlists.length, totalTracks, uid, nickname }; })() ``` ### Phase 2: Format Text ```javascript (() => { const d = window.__exportData; const m = window.__exportMeta; const entries = Object.entries(d).sort((a, b) => { if (a[1].specialType === 5) return -1; if (b[1].specialType === 5) return 1; return a[1].name.localeCompare(b[1].name, 'zh-CN'); }); const total = Object.values(d).reduce((s, p) => s + p.tracks.length, 0); const lines = [ `网易云音乐歌单导出`, `用户: ${m.nickname} (uid: ${m.uid})`, `导出时间: ${new Date().toLocaleString('zh-CN')}`, `歌单数: ${entries.length}`, `总歌曲数: ${total} (含重复)` ]; for (const [, pl] of entries) { lines.push('', '========================================', `歌单: ${pl.name}`, `歌曲数: ${pl.tracks.length}`, '========================================'); for (const t of pl.tracks) lines.push(`${t.artist} - ${t.name}`); } window.__musicTxt = lines.join('\n') + '\n'; return { chars: window.__musicTxt.length, lines: lines.length }; })() ``` ### Phase 3: Chunked Transfer Read in chunks of 40000 chars: ```javascript // Call N times, incrementing offset: 0, 40000, 80000, ... (offset) => { const chunk = window.__musicTxt.substring(offset, offset + 40000); return { offset, length: chunk.length, done: offset + 40000 >= window.__musicTxt.length, chunk }; } ``` Pass `offset` as the argument. Stop when `done === true`. ### Phase 4: Write to Disk Concatenate all chunks. Write via the `Write` tool or Python to the user's desired path. ## Output Format ``` 网易云音乐歌单导出 用户: <nickname> (uid: <uid>) 导出时间: <datetime> 歌单数: <N> 总歌曲数: <N> (含重复) ======================================== 歌单: <playlist_name> 歌曲数: <N> ======================================== <artist> - <song_name> <artist> - <song_name> ... ``` ## Rate Limits | Operation | Interval | |-----------|----------| | Playlist detail | 300ms between calls | | Song detail (batch) | 300ms, max 500 IDs per batch | | Total export time | ~30s for 30 playlists, ~60s for 1500+ songs | ## Edge Cases - **Subscribed playlists**: `n=2000` in playlist detail may only return 20 tracks. The trackIds array always has all IDs — use song/detail to fill gaps. - **Deleted/taken-down songs**: song/detail may return fewer songs than IDs requested. These are silently skipped. - **Empty playlists** (trackCount=0): Included in output with 0 songs, no tracks section. - **Special playlists**: `specialType=5` = liked songs ("我喜欢的音乐"), `specialType=100` = private radar, `specialType=300` = AI generated. ## Boundaries - Only exports the currently logged-in user's playlists - Does not export play counts, comments, or creation dates per song - Text format only (not CSV, JSON, or other structured formats unless user requests)