// Read and post to Twitter/X via the `bird` CLI — tweets, threads, search, timelines, bookmarks, follows, lists, media.
Ultra-compressed communication mode. Cuts tokens ~75%. Levels: lite, full, ultra, wenyan-lite/full/ultra. Trigger: /caveman, "be brief".
Write codebase docs: README, architecture, API, comments. Project structure and getting-started focus.
Compress memory files (CLAUDE.md, todos) to caveman format. Backup saved as FILE.original.md. Trigger: /caveman:compress <filepath>.
React/Preact frontend conventions: TanStack Query for server state, 8-section components, no TypeScript, no barrel files.
Manage Hetzner Cloud infrastructure via the hcloud CLI. Trigger: "hcloud", "hetzner", any Hetzner task.
Find architecture improvements: consolidate modules, improve testability and AI navigation. Trigger: improve architecture, refactor.
| name | bird |
| description | Read and post to Twitter/X via the `bird` CLI — tweets, threads, search, timelines, bookmarks, follows, lists, media. |
bird needs auth_token and ct0 from Firefox. node:sqlite is broken on this system, so extract manually before every bird session:
cp /home/mikekey/.mozilla/firefox/u80t1430.default-release/cookies.sqlite /tmp/ff_cookies.sqlite
sqlite3 /tmp/ff_cookies.sqlite "SELECT name, value FROM moz_cookies WHERE (host LIKE '%twitter.com%' OR host LIKE '%x.com%') AND name IN ('auth_token', 'ct0');"
Capture both values, then pass to every bird command:
bird <command> --auth-token <auth_token> --ct0 <ct0> --plain ...
Always use --plain (no emoji, no color — stable for LLM parsing). Use --json when structured data is needed. bird check verifies credentials without making other calls.
bird read <id-or-url> # single tweet (shortcut: bird <id-or-url>)
bird thread <id-or-url> # full conversation thread
bird replies <id-or-url> # replies (--all, --max-pages N)
bird search "query" # tweet search
bird mentions # tweets mentioning current user (or --user <handle>)
bird home # "For You" timeline
bird user-tweets <handle> # profile timeline
bird list-timeline <id-or-url> # tweets from a list
bird lists # your lists
bird news # AI-curated news from Explore (alias: trending)
bird bookmarks # your bookmarks
bird likes # your liked tweets
bird tweet "text" # post a tweet
bird reply <id-or-url> "text" # reply to a tweet
bird unbookmark <id-or-url>... # remove bookmarks (variadic)
bird follow <username-or-id> # follow
bird unfollow <username-or-id> # unfollow
bird following [handle] # users followed by current user (or handle)
bird followers [handle] # followers of current user (or handle)
bird about <username> # account origin + location info
bird whoami # verify which account the cookies belong to
--media <path> # repeatable, up to 4 images or 1 video
--alt "text" # alt text, repeated per media in order
Example:
bird tweet "caption" --media ./a.png --alt "a description" --media ./b.png --alt "b description"
bird check # verify credentials
bird query-ids # show/refresh cached GraphQL query IDs (--json)
--plain — stable, no emoji, no color. Default for LLM use.--json — structured output. Works on: read, replies, thread, search, mentions, bookmarks, likes, following, followers, about, lists, list-timeline, user-tweets, query-ids--json-full — adds raw API response in _raw field (tweet/reply only)--no-emoji, --no-color — individual overrides--quote-depth <N> — max quoted-tweet depth (default 1, 0 disables)--timeout <ms> — request timeoutauth_token and ct0 in shell variables--plain --auth-token "$AUTH_TOKEN" --ct0 "$CT0"bird reads ~/.config/bird/config.json5 and ./.birdrc.json5 (JSON5 format).
Supports: chromeProfile, chromeProfileDir, firefoxProfile, cookieSource, cookieTimeoutMs, timeoutMs, quoteDepth.
Env vars: NO_COLOR, BIRD_TIMEOUT_MS, BIRD_COOKIE_TIMEOUT_MS, BIRD_QUOTE_DEPTH.
/home/mikekey/.bun/bin/bird (v0.8.0)