
@viemccoy this does the job https://github.com/afar1/fieldtheory-cli
The workflow bypasses restricted Twitter APIs by mimicking human actions.
Many users praised an AI agent using Codex with computer use to organize Twitter bookmarks into a database, calling the idea genius because it turns forgotten links into searchable working memory they want to adopt immediately.
No Digg Deeper questions have been answered for this story yet.

@viemccoy this does the job https://github.com/afar1/fieldtheory-cli

@viemccoy I set this up as an Obsidian plugin.
Has a sync + categorization system and a couple pipelines to make specific categories more useful. Extracts recipes, maps locations, + summarizes news into a digest.
Working on getting it bundled up for others to use.

@viemccoy Initial version in 23 and 2024 was embeddings, then it became much more

@viemccoy like this? https://github.com/adamisom/best-tweets

@viemccoy I used the X API to generate a digest of my bookmarks and then clean them up. I now have a weekly automation running that generates a digest for me based on all my bookmarks and creates themeatic summaries with source citations from the original posts and content

@arestlessrest @viemccoy computer use is just overkill, run this script on the bookmarks page: https://x.com/626ripes/status/1760601684261396986 and it'll get all data out, then you can use the data however you want (including augment with LLMs)

@viemccoy

@viemccoy The fact that this is something we need computer use to do feels like neglect from X dot com.. let us pay for this, even

@viemccoy Done this for years in postgres, works amazing

@viemccoy wait this is actually genius
mine just sits untouched for eternity unfortunately

@viemccoy So true! I probably have hundreds of bookmarks by now, I wish I could just use a program to organize them by topic

@viemccoy you can do that with a chrome extension parsing everything in 2 seconds

@viemccoy Adding to my bookmarks to never see again

@viemccoy you can do the same with claude on chrome

@viemccoy use zotero you will want a web archive of tweets saved

@viemccoy I did this last year and everything lives in my Obsidian Vault. It’s been a complete game changer and I was always wondered why I never saw anyone mention this tactic on x.

@viemccoy I have branches on the field theory tool that do this for my Twitter, reddit, HN, GitHub stars, and my browser bookmarks.

@Ceres2079 @viemccoy i did this with Twitter bookmarks downloaded (chrome extension) → export to spreadsheet → upload to claude for tagging

@viemccoy bookmarks are underrated raw material. once they are tagged and searchable, they stop being a graveyard and become actual working memory.

@viemccoy Codex r n