x-twitter
Fetch X/Twitter posts, threads, replies, profiles, and search results via the FxTwitter (FxEmbed) API and mirror domains. Use whenever an x.com or twitter.com link appears, or the user asks about X/Twitter content. Direct scraping of x.com hits login walls and blocks.
- License 许可
- AGPL-3.0
- 最近更新
- 2026-11-04
- Stars
- ★ 18
- Forks
- 4
01 前言 Front Matter
- name
- x-twitter
- description
- Fetch X/Twitter posts, threads, replies, profiles, and search results via the FxTwitter (FxEmbed) API and mirror domains. Use whenever an x.com or twitter.com link appears, or the user asks about X/Twitter content. Direct scraping of x.com hits login walls and blocks.
03 SKILL.md 正文
X / Twitter access
Direct scraping of x.com / twitter.com fails (login walls, blocks). Two working paths, in order:
- FxTwitter REST API (preferred). Unofficial JSON API, no auth, read-only. Base:
https://api.fxtwitter.com - Mirror domains. For full-page rendering or when the API is down
Call endpoints with terminal_run + curl (or Python httpx/urllib).
API rules
- GET only. No API key, no login, no cookies. Guest public data only.
- Always send a
User-Agent(e.g.-A "tabyAgent/1.0"). Empty UA gets blocked. - Timeout 20s. Retry only on
429/502/503, max 3 tries, backoff 1s, 2s, 4s. - Every response JSON has a
codefield. Checkcode === 200, not just the HTTP status. HTTP 200 withcode !== 200is a failure. - ~1000 req/min per IP on paper; search gets throttled much earlier.
- Public posts only. Protected/suspended accounts and deleted posts return 401/404.
fxtwitter.comthe website is deprecated/gone; theapi.fxtwitter.comsubdomain still works.
Extract the tweet id
Supported inputs: bare numeric id, or a status URL on x.com, twitter.com, fxtwitter.com, fixupx.com, twittpr.com.
(?:x\.com|twitter\.com|fxtwitter\.com|fixupx\.com|twittpr\.com)/[^/]+/status/(\d+)
Capture group 1 is the tweet_id. The handle is not needed for lookups.
Endpoints
| Need | Call |
|---|---|
| Single post + metrics | GET /2/status/{id} |
| Author's thread (upward self-chain) | GET /2/thread/{id} |
| Thread + other users' replies | GET /2/conversation/{id}?ranking_mode=likes|recency |
| Keyword search | GET /2/search?q=...&feed=latest|top|media&count=1-100 |
| Profile | GET /2/profile/{handle}, handle or id:<numeric> |
| User timeline | GET /2/profile/{handle}/statuses?count=&cursor= |
| User media only | GET /2/profile/{handle}/media |
| Quotes / reposting users | GET /2/status/{id}/quotes · /2/status/{id}/reposts |
Common query params: lang (translation target, e.g. ko), cursor (from prior cursor.bottom), count (1–100), about_account=1.
curl -sS -A "tabyAgent/1.0" "https://api.fxtwitter.com/2/status/1548602399862013953?lang=ko"
curl -sS -A "tabyAgent/1.0" "https://api.fxtwitter.com/2/search?q=from%3Aelonmusk&feed=latest&count=20"
curl -sS -A "tabyAgent/1.0" "https://api.fxtwitter.com/2/conversation/1548602399862013953?ranking_mode=recency"
curl -sS -A "tabyAgent/1.0" "https://api.fxtwitter.com/2/profile/elonmusk"
Success is code: 200 plus the root key (status / results / user).
Useful status fields
id, url, text, created_at, created_timestamp, likes, reposts, quotes, replies, views (may be null), lang, author.{name,screen_name,avatar_url}, replying_to, quote (or tombstone), poll, media.photos[], media.videos[], article (X Articles).
Pagination & search notes
- If
cursor.bottomis present, request the same path with?cursor=. If it is null or empty, you are done. Keep to ~5 pages max. /2/conversationpage 2+ can return HTTP 404. Known bug; return what was collected and stop.- Search
qaccepts X web operators (from:,#tag,"exact phrase",min_faves:,since:). Best-effort, not guaranteed. - Search is recent/popular results, not a full archive. It fails more than single-post lookups: on
502/search_unavailablesay the search upstream failed, and fall back to/2/status/{id}if a URL/id is known. - User timeline extras:
with_replies(truthy),groupthreads,since(unix ts; 204 = nothing new). - Default to one page (~20 items); fetch more only if the user asks.
Routing
| User input | Endpoint |
|---|---|
| URL or numeric id | /2/status/{id} |
| "thread / 타래 / 스레드" | /2/thread/{id}, then /2/conversation if replies needed |
| "replies / 답글 / 댓글" | /2/conversation/{id} (ranking_mode=recency for chronological) |
| "search / 찾아 / 최근글" + keyword | /2/search |
@handle or "profile / 프로필" |
/2/profile/{handle} |
| "their posts / 그 계정 글" | /2/profile/{handle}/statuses |
Mirror domains (fallback / page view)
When the API can't help (need the rendered page, media embed, or API down), rewrite the host instead of browser-automating x.com:
- Replace
x.comwithfixupx.com(orxfixup.com) - Replace
twitter.comwithtwittpr.com - Further fallbacks:
fixvx.com,yt-dlp, or the browser tools on a mirror page. Try the next when one 404s/blocks.
Don'ts
- Never mix official
api.x.comschemas with this API. - Never mix the legacy
/{handle}/status/{id}form (root keytweet) with v2/2/...(root keystatus). Use v2 only. - No bulk search crawling; no cookies or account credentials. This API only serves guest-public data.
- Don't report success on
code !== 200even when HTTP is 200.
Error mapping
| code/HTTP | meaning |
|---|---|
| 400 | bad request |
| 401 | private or restricted |
| 404 | not found |
| 429 | rate limited |
| 500/502/503 | upstream error |
Limits: relay honestly to the user
- Unofficial proxy; can break or shut down without notice.
- Search is not a full archive; replies may be partial.
- Protected accounts and deleted posts are unreachable.
04 同仓库更多 Skills
xvfb
Linux GUI app automation on a virtual X display. Launch non-browser Linux GUI apps in Docker, see screenshots, and interact by pixel coordinates.
visualization
Deploy interactive HTML visualizations (quiz, what-if explorer, decision tree, or custom) to a live URL via Cloudflare Workers. Use when the user needs an interactive, clickable visual aid such as choices, buttons, sliders, causal "if this then that" explorers, or step-by-step decision flows. Not for plain charts or static images.
todos
The shared todo store for the user's tasks (suggest/approve only) and your own scheduled automations. Use for tasks, reminders, recurring checks, or "tell me when X".
skill-authoring
Create or update agent skills under the user skills directory. Use when a non-trivial, repeatable workflow just succeeded and should be captured for future sessions, or when the user asks to write or fix a skill.
privacy
Privacy tools and alternatives to mainstream services with verified details on licensing, ads, and pricing.
k-school
(한국 사용자 전용) 학교알리미와 컴시간을 이용하여 전국 학교 시간표와 급식을 가져옵니다. 이 스킬에는 브라우저를 이용한 웹 크롤링과 비공식 API 호출 방법 및 컴시간 난독화 우회 방법이 모두 포함되어 있습니다.