Twitter MCP (twikit)

This MCP server lets Claude search Twitter and pull timelines without paying for API access. It uses twikit under the hood, which means you authenticate with your actual Twitter account credentials rather than developer tokens. I've used it for quick competitor research and tracking niche discussions where the official API would cost more than the insight is worth. The main appeal is cost. If you're doing occasional Twitter research or need to grab a few hundred tweets for sentiment analysis, this sidesteps the monthly API bill entirely. You can search by keyword, pull user timelines, and grab tweet threads. It's particularly useful for archival workflows where you want to snapshot conversations before they disappear, or for tracking brand mentions without setting up a paid monitoring tool. The catch is that you're handing over your Twitter credentials. The repo stores them in environment variables, which is better than hardcoding, but you're still trusting the server process with login access. Twitter also rate-limits regular accounts more aggressively than API clients, so heavy scraping will get you temporarily blocked. I hit the limit after pulling about 300 tweets in quick succession. The server doesn't handle pagination gracefully either, you'll need to manually chunk large requests. Authentication can be fiddly. Twitter sometimes triggers email verification or CAPTCHA challenges when it sees automated login patterns, and the server doesn't handle those flows. You might need to log in manually via a browser first to clear any security flags. This is for developers who need sporadic Twitter data and don't want to justify an API subscription. If you're building production monitoring or need reliable high-volume access, pay for the real API. If you're doing one-off research, competitor tracking, or personal archiving, this does the job without the monthly cost. Just don't use your main Twitter account, create a dedicated research account instead.