dsh-x-archive

When an Agent needs to read an X (Twitter) post, scraping the page directly is fragile: images break, tweets get deleted or locked, and authors silently edit the text, leaving the Agent with stale or incomplete data. dsh-x-archive routes the post URL through the 2100laike archive service. Untweeted posts are automatically submitted for archiving, with images mirrored to a CDN so they survive deletion, and the original text stays accessible even after the tweet is removed. Already-archived posts return instantly from cache in full Markdown with zero quota cost, and tweets that were later edited can be re-fetched with a single refresh flag to track every version.

The design philosophy is simple: archive once, read forever. The Agent hands a tweet URL to the x_archive tool and receives structured Markdown body with metadata such as author, version, deletion status, and truncation flag. Overly long content is gracefully truncated with a snapshot link attached, so the model never loses context to an oversized response. No cookies, no login sessions, no scraping infrastructure, just a clean API call with an optional PAT for higher limits.

If you are running DeepSeek Harness agents that cite real-time social-media opinions, track how a viral tweet evolves over time, or compare content across days and weeks, dsh-x-archive gives you a reliable, persistent reading layer that a plain web scrape simply cannot match.