[timstark]

Every free method of getting Twitter data is dead or paid. Nitter shut down when Twitter killed guest accounts. Twint is broken. The official API starts at $100/month.

I found a combination that still works: DuckDuckGo indexes public tweets and returns URLs. fxtwitter (3,500+ stars, built for Discord embeds) has a public JSON API that returns full tweet data for any tweet ID — text, likes, retweets, replies, views, author, timestamps. No auth, no key, no account.

Nobody had combined the two. DuckDuckGo can search but can't read tweets. fxtwitter can read tweets but can't search. Together they give you a free, zero-credential Twitter search API.

  pip install trawlx
  trawlx --query "claude code" --mode json

That gives you ranked tweet objects as JSON to stdout. Pipe it wherever you want.

It also has a markdown digest mode and an optional claude -p integration for AI-summarized daily digests via launchd, but the core value is the data pipeline itself.

[redoxate]

Thanks for the share, were you able to quantify the delay vs real prod search ?

[timstark]

Thanks. Depends on DuckDuckGo's indexing speed — I've seen it pull tweets from within the last 30 minutes, but it varies. For my use case I'm pulling the top tweets from the past 24 hours as a daily digest, so the delta doesn't really matter.