🛠️ Twitter Archive Toolkit

Swiss army knife for experiments with my twitter archive.

Live demo: https://defenderofbasic.github.io/twitter-archive-toolkit/

The basic viewer defaults to loading my twitter JSON
You can replace it with your own JSON, either from the community-archive or locally
To get yours from the community-archive, see the URL list in data/bucket-urls.json
- To do it locally:
  - run the script ./scripts/raw-archive-to-json.js. Set the filepath to your twitter archive
  - this will combine all the relevant data from the ZIP file into one json
  - put that JSON into frontend/public/local-archives/ (or just host it anywhere)
  - give the app that URL

Semantic search

Clone this repo, run pnpm install
Run pnpm dev to launch frontend
In a separate terminal, run pnpm start-server
Open http://localhost:3000/semantic-search.html
Enter your OpenAI API key on the page, then refresh

This will automatically fetch the tweets from the given archive, embed them using OpenAI's embedding model, store the vector results in a local JSON file using Vectra. Once it's done embedding, you can search your tweets semantically.

DefenderOfBasic/twitter-archive-toolkit

🛠️ Twitter Archive Toolkit

Semantic search