birdx#

CLI for X/Twitter via direct API calls. Reads Chrome cookies from disk (no browser needed after birdx auth).

Install#

bash

bash <skill-dir>/scripts/install.sh

Or manually:

bash

npm install --prefix ~/clawd ws jsdom x-client-transaction-id
ln -sf <skill-dir>/scripts/birdx.js /opt/homebrew/bin/birdx
birdx auth

Auth#

bash

birdx auth          # Read Chrome cookies from disk (no browser needed)
birdx auth --cdp    # Use openclaw browser CDP fallback

Cookie cache: ~/.config/bird/birdx-cookies.json (TTL ~5h, re-run auth to refresh)

Commands#

bash

# Read single tweet (includes viewCount, bookmarkCount)
birdx read <tweet-url|id>

# Read thread + all replies (--sort relevance puts author thread first)
birdx replies <tweet-url|id> --sort relevance
birdx replies <tweet-url|id> --sort latest --total 40

# Search tweets (full metrics: views, likes, bookmarks, reposts)
birdx search "query" --sort top --total 20
birdx search "from:elonmusk" --sort latest

# Followers / following
birdx followers @username --total 50
birdx following @username --total 20

# JSON output
birdx read <url> --json
birdx search "query" --json

Output Fields#

• viewCount — total impressions
• likeCount, repostCount, replyCount, bookmarkCount
• author.userName, author.fullName, author.isVerified

Environment#

• BIRDX_MODULES — override node_modules path (default: $HOME/clawd/node_modules)
• Deps required: ws, jsdom, x-client-transaction-id
• Node 20+ required (uses node:sqlite)

Notes#

• Query ID cache auto-refreshes from X on first use. If stale: check ~/.config/bird/query-ids-cache.json
• birdx read may fail behind certain VPNs; birdx search is more reliable
• Bookmark count is only available via birdx (not bird CLI)