TESIGN
SAVED

EN · KO

TESIGN / RADAR

About TESIGN Radar

TESIGN Radar presents AI-selected daily tools and rising and all-time rankings, under the owner’s delegation. Stars, increases, cross-source signals, updates and licences are the evidence. AI chooses the lists under the owner’s delegation. No human review is running in September 2026. Total stars, increases, cross-source signals, last updates and licences are shown as evidence. Run by tesign, a one-person studio. No ads, paid placement, accounts or payments.

Ten tools is the rule for today’s selection. Rising ends at 100 and all-time at 250: every list is finite.

AI chooses the lists under the owner’s delegation. No human review is running in September 2026. Total stars, increases, cross-source signals, last updates and licences are shown as evidence.

First seen is a real system timestamp. Backfilled history is never presented as live discovery. Days without observations have hollow bars; numbers are not a hands-on review.

How ranks are computed

Rising requires at least 200 observed new stars in the selected window, or at least 100 total stars and an increase of 20% or more. Sort order is the window’s increase. Repositories created within 30 days get NEW. A 24-hour increase of at least 30% with only a GitHub signal gets SPIKE.

All-time rank = stars × survival × cross. Survival is 1.0 for a push within 90 days of the reference time, or 0.8 within 365 days. Cross is 1 + 0.1 × (distinct source count − 1), capped at 1.5. At least 1,000 stars are required; the library category is excluded. Only raw numbers and ranks are displayed.

Per-category all-time Top 20 applies the same formula within each category — within an item’s public category when the editor confirmed one, otherwise the classifier’s. No page is built for the library or other categories, nor for a category with fewer than three qualifying repositories. A repository with several categories appears in each of them.

Both rankings admit only repositories the AI classifier passed as usable products. Archived repositories, forks, filtered, discarded or merged products, star-farming name patterns (cracks, activators, miners), repositories without a README, link or tutorial collections (awesome-style lists), and libraries, frameworks, SDKs and runtimes are excluded by name, topic, description and a maintained list; misses are added as they are found. The reference time is the latest stored GitHub hour. Increases are collected observations, not a guarantee of every star event.

Total stars are the last-checked values plus observed growth through the reference time.

Milestones are observed events of the last 30 days. ★ 1k, 5k, 10k, 50k and 100k crossings are found by walking the reference-time star total backwards through the stored daily increases, so a day can be off in either direction: after the last checked total the walk undercounts and the true crossing is earlier, before it the walk overcounts and the true crossing is later or is not reported. Nothing is guessed before the history begins. A repository created within 30 days and first recorded by live ingestion within 72 hours of creation is noted with the hours between creation and that first sighting. The 7-day rising leader is one event per reference day. Only repositories that pass the ranking gates, plus already published items, are included.

Data cards are light pages generated for ranked repositories that have no edited entry: the union of the three rising windows, the all-time 250 and every category Top 20, minus edited entries, at most 600 pages. They carry only the stored GitHub description, README excerpt, licence and numbers — no editorial text.

Store hits are Apple’s public App Store top-100 free and paid charts for Korea and the United States, observed twice a day. Ranks, ratings and rating counts are Apple’s published figures; rising (up 20 or more places in 24 hours, or a rating count up 20% — at least +200 — in 7 days) and new entries (released within 30 days and charted within 7 days of release) follow fixed rules over two of our observations. No score of ours is involved. Ratings can be inflated by in-app review prompts, so the rating count and the rank change are shown next to them and nothing is ordered by rating alone. Today’s list admits at most two store apps, and only with a new-entry or rising signal.

The numbers page counts what this radar watches — repositories with stored star history, candidate products, ranking-eligible repositories, published items, the star / licence / language distributions, per-source and per-category counts and store observations. Every figure is a count of observations stored at build time, never a score of ours, and each block states its counting rule and reference time.

Radar in numbers →

Follow by RSS

Follow today’s list, the 7-day rising list and the milestones in an RSS reader. These static feeds are generated by the PC engine and include their build time.

FAQ →