Recurate

A place to search over all your bookmarks or images saved on X (twitter).

Bookmarks is specific for me but this can also be used in a more generalised way...

How does it work?

How to Use?

• Just go to any page on X (account, bookmarks, etc.) and use the scrape extension to scrape the tweets. Use this for setting up and using the extension.
• Clone the repository in your system.

git clone https://github.com/rohinish404/Recurate.git
cd Recurate

• Create a data folder inside src and paste both the files with names bookmarks.json and image_data.json.
• Create a venv for running the fastapi server as well as the scripts. (Advised to create in the root folder). Below are the commands for macos.

python3 -m venv venv
source venv/bin/activate

• Open the cloned repo in terminal and run the execute.sh file

./execute.sh

• Run the fastapi server. Port used in html scripts is 8000 so try to run the server on this port only.

fastapi run src/server/main.py

• Run the html scripts using serve command. (i use this command along with custom live reload because i don't use vscode :"))

npm install -g serve
cd src
serve

• Use the app 🎉

Note - Name the files as mentioned above only otherwise the app won't recognise these data files.

Demo

My.Movie.2.1.mov

Further improvements (that can be made but i won't be working on it until i feel like :))

• Both text/image search can be improved. For images something like - use a vlm to describe the image -> create embeddings of that description -> then create db. Maybe this can improve the search.
• Current scraping script doesn't take into account longer tweets (which has learn more..). So add that to script.
• Download images feature in memeSearch
• Open actual tweets by clicking somewhere on displayed tweet.

Thanks for trying out. Do star⭐️ the repo if you found this useful