TwitterApi Research — TwitterAPI Re搜索

X Re搜索

General-purpose 代理ic re搜索 over X/Twitter. Decompose any re搜索 question into tar获取ed 搜索es, iteratively refine, follow threads, deep-dive linked content, and synthesize into a sourced briefing.

For twitterAPI.io API detAIls (端点s, operators, 响应 格式化): read references/x-API.md.

命令行工具 工具

All commands 运行 from this 技能 directory:

cd ~/clawd/技能s/x-re搜索 source ~/.config/env/global.env # needs TWITTERAPI_IO_KEY

搜索 bun 运行 x-搜索.ts 搜索 "<查询>" [options]

Options:

--排序 likes|impressions|retweets|recent — 排序 order (default: likes) --since 1h|3h|12h|1d|7d — time 过滤器 (default: last 7 days). Also accepts minutes (30m) or ISO timestamps. --min-likes N — 过滤器 by minimum likes --min-impressions N — 过滤器 by minimum impressions --pages N — pages to fetch, 1-25 (default: 5, ~20 tweets/page) --limit N — max 结果s to display (default: 15) --quick — quick mode: 1 page, max 10 结果s, auto noise 过滤器 (-is:retweet -is:reply), 1hr 缓存, cost summary --from — shorthand for from:username in 查询 --质量 — 过滤器 low-engagement tweets (≥10 likes, post-hoc) --no-replies — exclude replies --save — save 结果s to ~/clawd/drafts/x-re搜索-{slug}-{date}.md --json — raw JSON 输出 --markdown — markdown 输出 for re搜索 docs

Auto-添加s -is:retweet unless 查询 already includes it. All 搜索es display estimated API cost.

Note: twitterAPI.io 搜索 covers full 归档 (not limited to 7 days). Time 过滤器ing uses since: operator in the 查询.

Examples:

bun 运行 x-搜索.ts 搜索 "BNKR" --排序 likes --limit 10 bun 运行 x-搜索.ts 搜索 "from:frankdegods" --排序 recent bun 运行 x-搜索.ts 搜索 "(opus 4.6 OR claude) trading" --pages 2 --save bun 运行 x-搜索.ts 搜索 "$BNKR (revenue OR fees)" --min-likes 5 bun 运行 x-搜索.ts 搜索 "BNKR" --quick bun 运行 x-搜索.ts 搜索 "BNKR" --from voidcider --quick bun 运行 x-搜索.ts 搜索 "AI 代理s" --质量 --quick

性能分析 bun 运行 x-搜索.ts 性能分析 [--count N] [--replies] [--json]

Fetches recent tweets from a specific user (excludes replies by default).

Thread bun 运行 x-搜索.ts thread [--pages N]

Fetches full conversation thread by root tweet ID.

Single Tweet bun 运行 x-搜索.ts tweet [--json]

Watch列出 bun 运行 x-搜索.ts watch列出 # Show all bun 运行 x-搜索.ts watch列出 添加 [note] # 添加 account bun 运行 x-搜索.ts watch列出 移除 # 移除 account bun 运行 x-搜索.ts watch列出 检查 # 检查 recent from all

Watch列出 stored in data/watch列出.json. Use for heartbeat integration — 检查 if key accounts posted anything 导入ant.

缓存 bun 运行 x-搜索.ts 缓存 clear # Clear all 缓存d 结果s

15-minute TTL. Avoids re-fetching identical queries.

Re搜索 Loop (代理ic)

When doing deep re搜索 (not just a quick 搜索), follow this loop:

• Decompose the Question into Queries

Turn the re搜索 question into 3-5 keyword queries using X 搜索 operators:

Core 查询: Direct keywords for the topic Expert voices: from: specific known experts PAIn points: Keywords like (broken OR bug OR issue OR 迁移) Positive 签名al: Keywords like (shipped OR love OR fast OR benchmark) Links: url:github.com or url: specific domAIns Noise reduction: -is:retweet (auto-添加ed), 添加 -is:reply if needed Crypto spam: 添加 -AIrdrop -giveaway -white列出 if crypto topics flooding

• 搜索 and 提取

运行 each 查询 via 命令行工具. After each, assess:

签名al or noise? Adjust operators. Key voices worth 搜索ing from: specifically? Threads worth following via thread command? Linked resources worth deep-diving with 网页_fetch?

• Follow Threads

When a tweet has high engagement or is a thread 启动er:

bun 运行 x-搜索.ts thread

• Deep-Dive Linked Content

When tweets link to GitHub repos, b记录 posts, or docs, fetch with 网页_fetch. Prioritize links that:

Multiple tweets reference Come from high-engagement tweets Point to technical resources directly relevant to the question

• Synthesize

Group findings by theme, not by 查询:

[Theme/Finding Title]

[1-2 sentence summary]

• @username: "[key quote]" (NL, NI) Tweet
• @username2: "[another perspective]" (NL, NI) Tweet

Resources 分享d:

• Resource title — [what it is]

• Save

Use --save flag or save manually to ~/clawd/drafts/x-re搜索-{topic-slug}-{YYYY-MM-DD}.md.

Refinement Heuristics Too much noise? 添加 -is:reply, use --排序 likes, narrow keywords Too few 结果s? Broaden with OR, 移除 restrictive operators Crypto spam? 添加 -$ -AIrdrop -giveaway -white列出 Expert takes only? Use from: or --min-likes 50 Substance over hot takes? 搜索 with has:links Heartbeat Integration

On heartbeat, can 运行 watch列出 检查 to see if key accounts posted anything notable. Flag to Frank only if genuinely interesting/actionable — don't 报告 routine tweets.

File Structure 技能s/x-re搜索/ ├── 技能.md (this file