Sliders are this book's suggestion budget. Saving locks allocation (the
desk will not Kelly-rebalance over your experiment). Resume auto when done.
Manual import โ JSON list, CSV, or Twitter archive tweets.js
โก Turbo mode โ feed the model your full history
The more history the model sees, the more it sounds like you and the
sharper its predictions. Two boosts, both free:
1 ยท Import your full X (Twitter) archive
On X: Settings โ Your account โ Download an archive of your data (takes ~24h to arrive).
Unzip it and locate data/tweets.js.
Drop that file below โ likes, reposts and replies come with each tweet. A large archive books on those stats (follower moves only exist for days the desk was running) and finishes scoring in the background.
2 ยท Import your Instagram export
Instagram: Settings โ Accounts Center โ Your information and permissions โ Download your information (choose JSON).
Drop your_instagram_activity/content/posts_1.json (and reels.json) below โ captions, timestamps and media formats (image / carousel / video) are parsed natively.
Drop tweets.js / JSON / CSV here, or click to choose
Also run Backfill follower history (below) so past posts can be
scored against real follower moves where the Internet Archive has snapshots of your profile.