Agentic Trading Lab
에이전트와 대화하고, 거래 아이디어를 테스트하며, 코드 없이 실제 시장 데이터에 AI 트레이딩 에이전트를 적용
아직 감지된 업데이트가 없습니다. GitHub 릴리스가 나오면 여기에 보입니다.
개인 계정의 공개 저장소 소유자는 GitHub로 로그인해 운영자를 확인하고 소개를 직접 관리할 수 있습니다. 관리자 승인은 필요하지 않습니다.
로그인한 뒤 확인 파일을 만들어 이 저장소의 기본 브랜치에 30분 안에 커밋해 주세요. 다른 브랜치나 포크의 파일은 인정하지 않습니다. 파일을 올린 뒤 아래에서 확인하면 관리자 승인 없이 관리할 수 있습니다. 자격을 유지하려면 파일을 남겨 두세요.
로그인이 필요하면 로그인 화면으로 이동합니다. 로그인 후 이곳으로 돌아와 원하는 확인 버튼을 다시 눌러 주세요.
프로젝트 폴더에서 /nomorevibe verify를 실행하면 소유권을 확인하고 소개를 직접 관리할 수 있습니다.
1분 만에 확인하기Build a block, it becomes a B-20 native block, paired with all market, have Fly brain, enjoy .
저장소 README에서 가져옴
README 전체를 불러오는 중입니다.
README 전문 보기 ↗DEFAULT_ADMIN_ROLE or MINT_ROLE — at any
point, creation included.0.This page holds no key, has no backend that can move funds, and presents transactions for your
wallet to sign. Every state it shows is re-read from the chain: a green result means a receipt with
status 0x1 was read back, never that a transaction was merely broadcast.
| moment | cost | where it goes |
|---|---|---|
| create a block and open its market | 0.001 ETH, once, in one signature | the fee wallet, in full |
| every swap afterwards | 0.5 % of the swap | the fee wallet, in full |
Fee wallet, and the only one: 0xa6cF99D35949c6cB911adB910078F4Ca46F0f5d4 — feeWallet() on the
hook returns it.
Opening a market used to cost 0.0003 ETH — not by choice. That was a floor inherited from an
older contract, and a $1 target in ETH falls below it at current prices, so the floor applied to
every launch, silently. Nobody had picked that price.
prix-du-lancement.mjs then read the 103 markets opened in one day by the busiest hook on Base:
43 distinct creators, median sent 0.001001 ETH. We were asking three times less than the
market, from the only side that actually pays. The new price is that median, rounded down.
⛔ This is not a revenue forecast. It aligns our price with the market's; it brings in no creator by itself. The effect will be measured — creations before and after — and published either way.
The hook declares address public immutable feeWallet. Written once, in the constructor. No
setter, no owner, no admin role of its own.
| can a fork change it | why | |
|---|---|---|
| where the fees land | no, on our hooks | immutable feeWallet, no setter |
| the 0.5 % swap rate | no, on our hooks | uint24 public constant HOOK_FEE = 5_000 |
| how much is sent at launch | yes | the hook only requires msg.value >= fraisVie, and fraisVie() reads 0.0003 ETH. The 0.001 ETH is our app's choice |
| all of it, by deploying its own hook | yes | nothing prevents it — DEPLOY.md has the recipe |
Editing FEE_WALLET in this repository changes what the app displays. It changes nothing on
chain, because the app does not route the fee — the hook does.
⛔ This section used to say a fork "cannot redirect the fees". That was an overclaim twice over,
and it is retracted rather than softened: a fork can deploy its own hook, and even on ours it can
send the on-chain minimum instead of what we ask — 70 % less. Measured on 2026-09-21 by calling
fraisVie(). test-frais-vont-au-wallet.mjs now fails if this page claims otherwise.
node test-frais-vont-au-wallet.mjs # reads feeWallet() on 4 deployed hooks + checks the source
node preuve-frais-arrivent.mjs # simulates a stranger opening a market, traces every transfer
The second one reports the path hop by hop rather than trusting a green status:
inconnu → hook : 1000000000000000 wei = 0.001000000 ETH
hook → a6cf : 1000000000000000 wei = 0.001000000 ETH
rien ne reste au hook : ✅ oui
DEPLOY.md explains how to make the fees go to you instead. We would rather write that down than have you discover it.
Read on 2026-09-21. Every figure below is reproducible with the scripts in this repository.
The hook, read on chain
| call | value |
|---|---|
HOOK_FEE() | 5000 / 1e6 = 0.5 % |
feeWallet() | 0xa6cF…f5d4 |
| fee split | 100 % to the fee wallet |
| currencies a block can be priced in | 17 — ETH, TBLOCK, USDC, cbBTC + 13 tokenized stocks |
The currency list lives in a deviseAdmise mapping written only in the constructor. It cannot be
extended after deployment — a constraint on us, and exactly what makes it trustworthy.
Fees collected — 14 days, 1818 log windows, zero missed
| fee collections | 41 (2.93 / day) |
| ETH reaching the fee wallet | 0.002798464 ETH |
| paid by addresses that are not ours | 32.8 %, from 5 distinct addresses |
Where B-20 volume actually lives — 1 day, 100 % of blocks read
| pools | swaps | distinct addresses | swaps per address | |
|---|---|---|---|---|
| pools with no hook | 20 | 282 | 19 | 4.3 |
| pools with a hook | 242 | 25 395 | 35 | 725.6 |
A hook does not repel traders — hooked pools get far more swaps. But 725 swaps per address is automation, not demand. Across the whole of Base, in one day, 54 distinct addresses touched a newly-born B-20 pool at all.
For scale — same day, same method
| token | distinct addresses | transfers |
|---|---|---|
| USDC | 329 365 | 3 610 687 |
| cbBTC | 10 816 | 583 892 |
| NVDAc | 2 158 | 35 403 |
| TBLOCK (ours) | 0 | 0 |
Published here because a page that only lists its wins is marketing, not evidence.
Every claim this file got wrong, kept next to what replaced it.
| what was said | what is true |
|---|---|
| "Nothing above has been done on Base mainnet" | 41 fee collections were measured over 14 days |
| "Real Create charges a 0.001 ETH service fee" | utiliseCreateRouter() returns false; Create went to the factory at value zero |
| "Create is free" | true of one call, false for anyone reading it — opening a market cost 0.0003 ETH |
| "B-20 tokens do get traded, at scale" | 43 263 swaps looked like a crowd; the address column said 35 addresses |
"main holds the retired GitHub Pages stub" | Pages reports status: built and is still serving |
| "a fork cannot redirect the fees" | it can, two ways: its own hook, or sending the on-chain minimum (0.0003 ETH) instead of the 0.001 our app asks |
Each script states, in its own header, the bound of what it can prove — and several refuse to give a verdict rather than publish a biased one.
node devises-admises.mjs # what a deployed hook accepts as a pair currency
node ou-vit-le-volume.mjs 1 # hooked pools vs free pools, by swaps AND by addresses
node bridge-devises-liquidite.mjs # how much each admitted currency actually moves
node prix-du-lancement.mjs # what creators pay to open a market elsewhere on Base
node reference-frais.mjs # fees now, against a reference frozen before the V6 hook existed
for f in test-*.mjs; do node "$f"; done # 37 files, all green on 2026-09-21
Two of these instruments were rebuilt after being caught measuring badly:
ou-vit-le-volume.mjs first read 19 windows out of 303 and would have published a ratio built
on 6 % of the period — and the windows that failed were the busiest ones, so the bias ran toward
the conclusion. It now splits a window in half on failure and refuses any verdict if a single
block stayed unread.test-texte-a-l-ecran.mjs first scanned one page while the defect it was written for sat on four.Static files. No build step, no dependencies, no backend.
git clone https://github.com/Tokenized-Block/tokenizedblock-app.git
cd tokenizedblock-app
python -m http.server 8000
Then open http://localhost:8000/app.html.
⚠️ A wallet's in-app browser on Android refuses plain HTTP (ERR_CLEARTEXT_NOT_PERMITTED). Use the
hosted HTTPS page on mobile, or a desktop browser with a wallet extension.
No undated figures. No ARR, MRR or projection appears anywhere in this repository. A number that was not measured is not written down, and a measurement that went against us is printed next to one that did not.