bsky-follows-backend

Implementation Plan Here's the comprehensive plan to add all the requested features:

1. Blacklist & Whitelist System

• Hardcoded blacklist: Any handle ending with bsky.app
• Whitelist: pfrazee.com (allowed despite being in top500)
• Dynamic blacklist: Sync from https://vqv.app/top500.json once per day
  • Filter out entries where is_ad: true
  • Exclude pfrazee.com
  • Store in SQLite blacklist table

2. Compression for Large Accounts

• Only cache accounts with followers >= 5000 OR follows >= 5000
• Use pako (gzip) to compress JSON before storing
• Decompress when retrieving from cache
• This allows ~70-85% size reduction

3. Database Schema -- Existing cache table (no changes needed) CREATE TABLE IF NOT EXISTS bsky_cache ( handle TEXT PRIMARY KEY, followData TEXT, -- Now compressed with gzip cacheHits INTEGER DEFAULT 0, lastCached INTEGER ); -- New blacklist table CREATE TABLE IF NOT EXISTS blacklist ( handle TEXT PRIMARY KEY, addedAt INTEGER, source TEXT -- 'top500' or 'manual' ); -- New sync tracking table CREATE TABLE IF NOT EXISTS sync_log ( syncType TEXT PRIMARY KEY, lastSynced INTEGER );
4. Request Flow
5. Check if handle ends with bsky.app → REJECT (403)
6. Check if handle in blacklist table → REJECT (403)
   • Unless handle is in whitelist (pfrazee.com)
7. Auto-sync top500 blacklist if >24hrs since last sync
8. Fetch profile to get follower/following counts
9. If counts < 5000 → Return fresh data (no caching)
10. If counts >= 5000:
    • Check cache
    • If cached → decompress & return
    • If not → fetch, compress, store, return
11. Key Functions to Add

• isBlacklisted(handle) - Check against all blacklist rules
• syncTop500Blacklist() - Fetch & update blacklist table
• shouldSyncBlacklist() - Check if >24hrs since last sync
• compressData(data) - Gzip compression using pako
• decompressData(compressed) - Gzip decompression
• shouldCache(followers, follows) - Check if >= 5000 threshold

6. Error Messages { error: Access denied, message: This account is not available for querying, status: 403 } Benefits:

• Protects your API from expensive queries on mega-accounts
• Caches strategically only large accounts that need it
• Saves storage with 70-85% compression
• Auto-updates blacklist daily without manual intervention
• Capacity: ~1,200-1,500 large cached accounts in 1GB Ready to implement when you are!