ramona.easton

ramona.easton

I like turning awkward chores into small scripts. Still figuring it out as I go.

Joined 10/30/2026
2
Projects
0
Collaboration requests
0
Iterated
0
Likes
0
Saves
41
Views
0
Comments
Content & Mediaclaude-code

麦田 Maitian

A one-person publishing desk for video creators: fill in seven platforms once, watch the numbers roll in, and see which episode actually paid.

One person making videos has to post to six or seven platforms, and each one wants different copy, a different cover and a different format. Then you still wonder which episode grew followers and which one brought paying customers. Maitian is the workbench its author, a Chinese creator who teaches AI courses, built for his own daily publishing and still uses every day. It works in three parts. A publishing skill lets an AI assistant watch the finished video, write each platform's copy and pack a release. A Chrome extension fills in each platform's upload page in batches. For YouTube, X, Instagram and your own site it checks the text and timing, then submits; for Bilibili, Douyin, Xiaohongshu and WeChat Channels it stops at filled-in, and you click publish yourself. The numbers side is just as thorough. Every episode's plays and interactions on every platform, follower counts, a big-screen dashboard on your phone, a card in your Feishu group each day, a review card on day 1, 3 and 7 that compares the episode with the previous one at the same age, and a funnel from plays to sign-ups to revenue if you connect your own site. It has 1,644 stars on GitHub with 483 forks, was created on 14 November 2026 and was last pushed on 20 November 2026. It is MIT licensed. It needs a Mac that stays on, Node.js 22.5 or newer and Chrome — a community Windows version exists, without autostart yet. It never stores any platform password and only reads your own creator accounts: no auto-liking, no touching other people's accounts.

Health & Fitnessclaude-code

Sports Video Coach

Turn one phone video of your practice into a coaching page with drills, clips and an explained video, built to work hand-in-hand with your coding agent.

Sports Video Coach is for players who film their own practice and never watch it back. You hand it one phone video, and it turns it into a coaching page you can open on your phone: what to fix, a drill for each problem, and the clips that show it, plus a spoken version you can listen to on the way to practice. The clever part is how it works with your coding agent instead of pretending to be one. The tool pulls frames, cuts clips and builds the page, then pauses and tells the agent exactly what to look at or write next. The agent reviews the frames, writes the review, and the run carries on until it finishes. You can see what you get without installing anything: the demo page is a synthetic drill with a pre-written review, and the explained video pauses where a shot goes wrong, rings the spot and subtitles why. The project is open source under MIT and has 2 GitHub stars. Honest limits: the demo shows the output format, not how good the agent is at basketball — that takes your own video. The agent judges frame sheets, not the video itself, so very fast details like the moment of release can fall between frames. Narration goes through an online text-to-speech service unless you skip audio, and model use may cost money.