X Twitter Scraper

Name: X Twitter Scraper
Author: xquik

Remote MCP server for X (Twitter) with 121 API endpoints via 2 tools. Search tweets, look up users, post tweets, like, retweet, follow, send DMs, run giveaway draws, monitor accounts, extract bulk data, compose AI-optimized tweets, and more.

该来源不提供完整文件导出（国内平台多为平台内托管），仅存元数据与原链

接入信息

传输形态: http
鉴权方式: 鉴权未知
端点: https://x-twitter-scraper.run.tools

鉴权方式未标注，请核对官方文档后再接入——不要直接使用以下片段

{
  "mcpServers": {
    "X Twitter Scraper": {
      "url": "https://x-twitter-scraper.run.tools"
    }
  }
}

能力清单

工具

说明

explore

Search the authenticated Xquik endpoint catalog without making network calls. Provide an async arrow function over spec.endpoints (EndpointInfo[]). Each endpoint includes method, path, summary, category (account, composition, credits, extraction, media, monitoring, support, twitter, x-accounts, x-write), access metadata, parameters, and responseShape. Filter to the smallest useful result. Examples: - `async () => spec.endpoints.filter(e => e.category === 'composition')` - `async () => spec.endpoints.filter(e => e.summary.toLowerCase().includes('tweet'))` - `async () => spec.endpoints.find(e => e.path === '/api/v1/x/tweets/search')` Read-only and idempotent. Requires MCP authentication to execute. Invalid or throwing code returns an error.

xquik

Execute authenticated Xquik REST API calls across 120 catalog routes. Binary responses are unavailable through MCP. Use their REST download endpoints. Provide an async arrow function using `xquik.request(path, { method?, body?, query? })`. Authentication is injected. The sandbox has no filesystem or general network access. Calls use normalized v1 responses. Pagination controls use snake_case. Resource fields follow endpoint schemas. Errors are structured. Timestamps use Unix seconds. Dependencies fail with HTTP 424. POST and DELETE operations may modify data. Prefer `options.query`. Inline query strings are accepted for compatibility. When both provide the same key, `options.query` wins. Discover unfamiliar paths with explore. Common routes: - Search: `async () => xquik.request('/api/v1/x/tweets/search', { query: { q: 'AI agents', limit: '100' } })` - Timeline: `async () => xquik.request('/api/v1/x/users/xquik/tweets')` - Current topics: `async () => xquik.request('/api/v1/radar')`. Select reddit, github, trustmrr, hacker_news, google_trends, wikipedia, or polymarket as the source. For reddit, use title, url, published_at, and metadata fields author, subreddit, selftext, content_url, score, upvote_ratio, estimated_upvotes, estimated_downvotes, number_comments, gallery_image_urls, and reddit_video. Treat estimated vote counts as estimates. The number_comments field is only a count, not comment text. Pass selected facts into compose as additionalContext. For trustmrr, always include metadata.x_handle as a plain @mention when present. This helps the founder find the post and may lead to replies, quotes, or reposts. - Draft text: pass selected Radar facts into POST /api/v1/compose as additionalContext. Run compose, refine, then score. - Publish: get an account from GET /api/v1/x/accounts, then POST /api/v1/x/tweets. Use numeric user IDs for follow and DM paths. - Extractions: POST /api/v1/extractions/estimate, POST /api/v1/extractions, then poll GET /api/v1/extractions/:id. A 202 response means accepted, not completed. Paginate while has_more is true. Pass next_cursor as cursor for most X reads, after for draws, extractions, events, and radar, and afterCursor for drafts. Page sizes and limits are upper bounds. For high-volume work, aggregate inside the sandbox or project only the fields the user needs so the result stays within the response limit. Read APIs preserve every safe field X provides. GET /api/v1/x/tweets/<tweet_id>/replies?mode=complete&limit=25000 merges available timeline views, supported rankings, every forward cursor module, labeled hidden-content branches, parent windows, and search. Direct: inReplyToId. Exclude nested_replies. Complete needs 80% of current reported direct replies. HTTP 424 replies_incomplete returns diagnostics. Inspect reportedReplyCount, uniqueDirectReplies, coveragePercentage, strategiesAttempted, cursorFailures, missingResponseModulesOrFields, and recommendedFallback. Trust diagnostic.complete. Use structured error codes. Fix request-shape errors such as missing_query, missing_ids, too_many_ids, missing_params, invalid_username, and unsupported_field before retrying. Respect retry_after and error.retryable. Confirm identifiers after user_not_found, invalid_tweet_id, or article_not_found. Resolve account_required or dm_not_permitted with a valid connected account. Replace an unreachable URL after media_download_failed. Verify state before retrying x_write_ambiguous. Report no_credits, insufficient_credits, no_subscription, or subscription_inactive. Inspect payment_options when present. A failed API request never creates checkout. Never call a billing POST without an explicit user request or confirmation. Ask the user to select the amount and payment option. After confirmation, generated code may use only an account credit or subscription create_checkout path present in this executable catalog, with its body and required_headers. Return the URL named by response_url_field. guest_wallet actions are direct REST only. Never call /api/v1/guest-wallets, /api/v1/guest-wallets/topups, or /api/v1/guest-wallets/status through xquik generated code. After explicit confirmation, report that direct REST POST /api/v1/guest-wallets accepts $10-$250 USD with a cryptographically random UUID v4 Idempotency-Key and returns a raw Stripe-hosted checkout_url, api_key, status_url, and poll_after_seconds. Give only checkout_url to the user. The direct REST caller stores api_key and the Idempotency-Key as secrets, then polls status_url after the stated delay. A paid_reads guest key uses direct REST POST /api/v1/guest-wallets/topups with the same api_key after confirmation. Do not include API keys or authorization headers in code. Return compact summaries or projections instead of full raw datasets when possible.

纠错与举报（发现条目失效、署名有误或涉及侵权？）

提交举报 / 纠错

侵权举报经核验成立后，我们会即时下线该条目并删除已存的内容副本。