1#!/usr/bin/env npx tsx
2 
3import * as path from "node:path";
4import { fileURLToPath } from "node:url";
5import { hasRequiredXCookies, loadXCookies } from "./cookies.js";
6import { fetchTweetThread } from "./thread.js";
7import { formatArticleMarkdown } from "./markdown.js";
8import { resolveReferencedTweetsFromArticle } from "./referenced-tweets.js";
9import { formatThreadTweetsMarkdown } from "./thread-markdown.js";
10import { resolveArticleEntityFromTweet } from "./tweet-article.js";
11 
12type TweetToMarkdownOptions = {
13  log?: (message: string) => void;
14};
15 
16function parseArgs(): { url?: string } {
17  const args = process.argv.slice(2);
18  let url: string | undefined;
19 
20  for (const arg of args) {
21    if (!arg.startsWith("-") && !url) {
22      url = arg;
23    }
24  }
25 
26  return { url };
27}
28 
29function normalizeInputUrl(input: string): string {
30  const trimmed = input.trim();
31  if (!trimmed) return "";
32  try {
33    return new URL(trimmed).toString();
34  } catch {
35    return trimmed;
36  }
37}
38 
39function formatScriptCommand(fallback: string): string {
40  const raw = process.argv[1];
41  const displayPath = raw
42    ? (() => {
43        const relative = path.relative(process.cwd(), raw);
44        return relative && !relative.startsWith("..") ? relative : raw;
45      })()
46    : fallback;
47  const quotedPath = displayPath.includes(" ")
48    ? `"${displayPath.replace(/"/g, '\\"')}"`
49    : displayPath;
50  return `npx -y bun ${quotedPath}`;
51}
52 
53function parseTweetId(input: string): string | null {
54  const trimmed = input.trim();
55  if (!trimmed) return null;
56  if (/^\d+$/.test(trimmed)) return trimmed;
57 
58  try {
59    const parsed = new URL(trimmed);
60    const match = parsed.pathname.match(/\/status(?:es)?\/(\d+)/);
61    if (match?.[1]) return match[1];
62  } catch {
63    return null;
64  }
65 
66  return null;
67}
68 
69function buildTweetUrl(username: string | undefined, tweetId: string | undefined): string | null {
70  if (!tweetId) return null;
71  if (username) {
72    return `https://x.com/${username}/status/${tweetId}`;
73  }
74  return `https://x.com/i/web/status/${tweetId}`;
75}
76 
77function formatMetaMarkdown(meta: Record<string, string | number | null | undefined>): string {
78  const lines = ["---"];
79  for (const [key, value] of Object.entries(meta)) {
80    if (value === undefined || value === null || value === "") continue;
81    if (typeof value === "number") {
82      lines.push(`${key}: ${value}`);
83    } else {
84      lines.push(`${key}: ${JSON.stringify(value)}`);
85    }
86  }
87  lines.push("---");
88  return lines.join("\n");
89}
90 
91function extractTweetText(tweet: any): string {
92  const noteText = tweet?.note_tweet?.note_tweet_results?.result?.text;
93  const legacyText = tweet?.legacy?.full_text ?? tweet?.legacy?.text ?? "";
94  return (noteText ?? legacyText ?? "").trim();
95}
96 
97function isOnlyUrl(text: string): boolean {
98  const trimmed = text.trim();
99  if (!trimmed) return true;
100  return /^https?:\/\/\S+$/.test(trimmed);
101}
102 
103export async function tweetToMarkdown(
104  inputUrl: string,
105  options: TweetToMarkdownOptions = {}
106): Promise<string> {
107  const normalizedUrl = normalizeInputUrl(inputUrl);
108  const tweetId = parseTweetId(normalizedUrl);
109  if (!tweetId) {
110    throw new Error("Invalid tweet url. Example: https://x.com/<user>/status/<tweet_id>");
111  }
112 
113  const log = options.log ?? (() => {});
114  log("[tweet-to-markdown] Loading cookies...");
115  const cookieMap = await loadXCookies(log);
116  if (!hasRequiredXCookies(cookieMap)) {
117    throw new Error("Missing auth cookies. Provide X_AUTH_TOKEN and X_CT0 or log in via Chrome.");
118  }
119 
120  log(`[tweet-to-markdown] Fetching thread for ${tweetId}...`);
121  const thread = await fetchTweetThread(tweetId, cookieMap);
122  if (!thread) {
123    throw new Error("Failed to fetch thread.");
124  }
125 
126  const tweets = thread.tweets ?? [];
127  if (tweets.length === 0) {
128    throw new Error("No tweets found in thread.");
129  }
130 
131  const firstTweet = tweets[0] as any;
132  const user = thread.user ?? firstTweet?.core?.user_results?.result?.legacy;
133  const username = user?.screen_name;
134  const name = user?.name;
135  const author =
136    username && name ? `${name} (@${username})` : username ? `@${username}` : name ?? null;
137  const authorUrl = username ? `https://x.com/${username}` : undefined;
138  const requestedUrl = normalizedUrl || buildTweetUrl(username, tweetId) || inputUrl.trim();
139  const rootUrl = buildTweetUrl(username, thread.rootId ?? tweetId) ?? requestedUrl;
140 
141  const articleEntity = await resolveArticleEntityFromTweet(firstTweet, cookieMap);
142  let coverImage: string | null = null;
143  let remainingTweets = tweets;
144  const parts: string[] = [];
145 
146  if (articleEntity) {
147    const referencedTweets = await resolveReferencedTweetsFromArticle(articleEntity, cookieMap, { log });
148    const articleResult = formatArticleMarkdown(articleEntity, { referencedTweets });
149    coverImage = articleResult.coverUrl;
150    const articleMarkdown = articleResult.markdown.trimEnd();
151    if (articleMarkdown) {
152      parts.push(articleMarkdown);
153      const firstTweetText = extractTweetText(firstTweet);
154      if (isOnlyUrl(firstTweetText)) {
155        remainingTweets = tweets.slice(1);
156      }
157    }
158  }
159 
160  const meta = formatMetaMarkdown({
161    url: rootUrl,
162    requestedUrl: requestedUrl,
163    author,
164    authorName: name ?? null,
165    authorUsername: username ?? null,
166    authorUrl: authorUrl ?? null,
167    tweetCount: thread.totalTweets ?? tweets.length,
168    coverImage,
169  });
170 
171  parts.unshift(meta);
172 
173  if (remainingTweets.length > 0) {
174    const hasArticle = parts.length > 1;
175    if (hasArticle) {
176      parts.push("## Thread");
177    }
178    const tweetMarkdown = formatThreadTweetsMarkdown(remainingTweets, {
179      username,
180      headingLevel: hasArticle ? 3 : 2,
181      startIndex: 1,
182      includeTweetUrls: true,
183    });
184    if (tweetMarkdown) {
185      parts.push(tweetMarkdown);
186    }
187  }
188 
189  return parts.join("\n\n").trimEnd();
190}
191 
192async function main() {
193  const { url } = parseArgs();
194  if (!url) {
195    console.error("Usage:");
196    console.error(`  ${formatScriptCommand("scripts/tweet-to-markdown.ts")} <tweet url>`);
197    process.exit(1);
198  }
199 
200  const markdown = await tweetToMarkdown(url, { log: console.log });
201  console.log(markdown);
202}
203 
204const isCliExecution =
205  process.argv[1] && fileURLToPath(import.meta.url) === path.resolve(process.argv[1]);
206 
207if (isCliExecution) {
208  main().catch((error) => {
209    console.error(error instanceof Error ? error.message : error);
210    process.exit(1);
211  });
212}
213