On-chain signals
How a young coin is really trading: who is buying and selling, counted in wallets, read as one setup with a 0–100 score and how traders usually play it.
Price and volume are easy to push. Wallets are harder to fake. On-chain signals reads what the wallets trading a young coin are doing, and sums it up as one setup.
Open any coin and the Signals tab shows the read. In Glimz AI, ask "how is it trading" or type /signals <coin>. Glimz AI also runs it next to Buy or not for every coin under a few days old.
The signals
Each signal compares a number with a fixed threshold. Green helps, red hurts, plain is neutral.
| Signal | What it measures | Reads as a warning when |
|---|---|---|
| Unique buyers (1h) | Different wallets that bought in the last hour | Under 25 (a crowd starts at 150) |
| Buyers vs sellers (wallets, 1h) | Buying wallets ÷ selling wallets | 0.85 to 1 or less (good at 1.3 to 1 or more) |
| Sells per selling wallet (1h) | How often each seller sold | 6 or more, with at least 80 sells, while buyers buy once or twice |
| Trades per wallet (1h) | Trades ÷ wallets | 10 or more, or buyers and sellers both trading 4 or more times each, with at least 300 trades |
| Average trade (1h) | Volume ÷ trades | Under $8, with at least 300 trades |
| Busiest 3 wallets' share | Share of the latest trades made by 3 wallets | 40% or more |
| Volume vs liquidity (1h) | Hourly volume ÷ pool liquidity | 8x or more (active from 0.5x) |
| Buying pace now vs the hour | Buyers per minute in the last 5 minutes ÷ the hour's average | 0.4x or less (heating up at 1.5x) |
| Market cap vs liquidity | How thin the pool is for its size | 30x or more |
| Top 10 holders | Share of supply in the 10 biggest wallets | 45% or more |
| Developer holds | The developer wallet's share | 10% or more |
| Mint / freeze authority (Solana) | Whether someone can print or freeze the coin | Still set |
| Trades of $1,000+ | Big buys vs big sells in the latest trades | More big sells than big buys |
| Wallets that passed the copy check | Wallets from the Copy radar in the latest trades | More of them selling than buying |
Signals only show when there is enough data for them. A coin still on a launchpad curve has no pool liquidity yet, so there's no liquidity signal; the curve's progress shows instead.
The setups
The signals add up to one setup, checked in this order. The first that fits wins.
| Setup | When | What it means |
|---|---|---|
| Few wallets selling | Sells per seller is high while buyers buy once | A small group is selling over and over into buyers. Usually the developer or early insiders cashing out. |
| Bot volume | Many trades per wallet, buyers and sellers both repeating, tiny trades, or 3 wallets making most trades | The volume isn't real interest. |
| Selling off | Down 30% or more in an hour, or more selling wallets than buying ones with the price falling | No support yet. |
| Too quiet to read | Under 30 trades in the hour | Nothing to read. |
| Just launched | Under an hour old | Snipers and first buyers; the crowd hasn't formed. |
| Already pumped | Up 100% or more in an hour | The easy part of the move is over. |
| Launch-curve run | Still on a launchpad curve with buyers arriving | A bet on graduating to a real pool. |
| Early momentum, broad buyers | 150+ unique buyers, 1.3+ buyers per seller, buying not slowing | A real crowd is buying. |
| Steady buying | More buyers than sellers, 3+ hours old, up but not pumped over 6 hours | Slower but healthier than a pump. |
| Mixed signals | Anything else | Nothing stands out. |
The signals score (0–100) sums the signals. It says how healthy the activity looks for a young coin. It is not a price prediction.
Coin age
Age counts from the coin's first pool anywhere, so a new pool for an old coin doesn't pass for a new coin. The read is built for coins under 3 days old; older coins get a note.
How traders usually play each setup
Every read comes with how traders usually play that setup and what would change the read. For example, for Early momentum: keep the stake small, wait for the first pullback to hold rather than buying a green candle, and take the stake back out at 2x. See the playbook for the full picture.