Analyze the content strategy of X/Twitter bloggers and extract creation methods worth learning from.
⚠️ Important limitations
X/Twitter uses Cloudflare human verification for protection,the success rate of automatic scraping is very low. The following workflow is recommended:
Recommended workflow (manual data + AI analysis)
用户提供博主链接 → 用户手动复制推文 → AI 分析 → 生成报告 → 保存到 05 笔记/01 博主分析/
How to operate:
- The user opens the blogger's page in the browser
- Scroll to load 20-50 tweets
- Select all and copy the tweet content (Cmd+A, Cmd+C)
- Paste it to the AI for analysis
Automatic scraping workflow (low success rate)
用户输入博主链接 → 提取用户名 → 尝试抓取推文 → AI 分析 → 生成报告
Step 1: Parse the blogger link
Extract the username from the link:
https://x.com/elonmusk→elonmuskhttps://twitter.com/elonmusk→elonmusk
Step 2: Obtain tweet content
Method A: User provides manually (recommended)
Ask the user to:
- Open in the browser
https://x.com/{username} - Scroll the page to load more tweets
- Copy the tweet content and paste it into the conversation
Prompt:
Because X/Twitter has anti-scraping protection, please open in your browser https://x.com/{username}，滚动加载 after 20-50 tweets, copy the content and paste it to me, and I will help you analyze it.
Method B: Automatic scraping (may fail)
Run the scraping script:
python3 scripts/fetch_tweets.py <username> [--count 100]
Common reasons for failure:
- Blocked by Cloudflare human verification
- Login is required to view
- IP temporarily banned
Method C: Third-party mirror sites
Try a nitter mirror site (may be unstable):
https://nitter.net/{username}https://nitter.cz/{username}
Step 3: AI analysis
Based on the obtained content, analyze the following dimensions:
3.1 Analysis of reasons for viral hits
- What is the core driving force behind follower growth?
- What common characteristics do viral contents have?
- Which tweets received unusually high engagement?
3.2 Content style analysis
- Language style: formal/colloquial, humorous/serious, long sentences/short sentences
- Topic area: What topics are mainly discussed? Is there a fixed theme?
- Expression method: narrative/opinion/tutorial/interactive
3.3 Creation frequency analysis
- Average daily/weekly posting volume
- Active time periods (morning/noon/evening/late night)
- Whether there is a fixed update rhythm
3.4 Creation method analysis
- Original vs repost ratio
- Long-form vs short-form ratio
- Whether images/videos are included
- Usage of Thread long tweets
3.5 Strategies worth learning from
- Which practices are worth learning from?
- How can they be applied to one's own creation?
- What problems need to be avoided?
Step 4: Generate report
Use the references/report_template.md template to generate an analysis report.
Save location: 05 笔记/01 博主分析/
Naming convention: MMDD博主分析-用户名.md(such as 0111博主分析-elonmusk.md）
Script description
scripts/fetch_tweets.py
Use Playwright to scrape tweet content. Dependencies need to be installed first:
pip install playwright
playwright install chromium
Note: Due to Cloudflare protection, this script has a low success rate.
references/report_template.md
Markdown template for the analysis report, containing the structured output format for all analysis dimensions.
Frequently Asked Questions
Q: What should I do if automatic scraping fails? A: This is normal; X/Twitter has strict Cloudflare protection. Please use the manual copy method:
- Open the blogger's page in a browser
- Copy the tweet content
- Paste it to AI for analysis
Q: How can I get more tweets? A: Keep scrolling the page in the browser to let more tweets load, then select all and copy.
Q: Can I analyze other platforms? A: If the blogger has accounts on other platforms such as Jike, Weibo, or WeChat Official Accounts, their content can also be provided for analysis.
Q: The analysis results are inaccurate? A: The more tweets there are, the more accurate the analysis. It is recommended to provide at least 20 tweets.