How Kairos reads a trench while it forms
Kairos does not ask what pumped. It asks who already called a new token, at which minute, and how that person's past calls ended.
Every new Pump.fun token is matched against a growing database of shillers and their history. Within minutes of the first post with the token's contract address, Kairos produces one signal: who is in, when they came in, which past cases it resembles and how confident the system is.
How it works
Five steps run for every token, from launch to the signal in Telegram.
Data sources
| Source | What Kairos takes from it |
|---|---|
| PumpPortal | New Pump.fun tokens and migrations as they happen. |
| Frontrun | Posts that contain a contract address, the accounts behind them, follower quality (SMARTS) and deleted posts. |
| Helius | On-chain trades and token history on Solana, used to label how each token ended. |
The database is what matters. It records shiller → token → outcome every day, so each new signal rests on a longer history. It cannot be rebuilt quickly by someone starting today.
Signal card
The card is what you receive within minutes of the first shill. Each field answers one question.
| Field | What it tells you |
|---|---|
| CA | The token's contract address. |
| Age | Minutes since the token was created. |
| MCAP | Market cap in SOL at the moment of the signal. |
| Holders | Number of wallets holding the token. |
| Timeline | Minutes since creation, with the minute each shiller posted. |
| Who is already in | Shillers from the database who posted the CA, with followers, number of calls, SMARTS, CA count, win rate and rug rate. |
| Looks like | The most similar past tokens and how they ended, with a similarity percentage. |
| Confidence | A 0–100 score from rules and statistics. |
| Risks | Plain warnings, for example a creator who launched several tokens this week, or liquidity still on the bonding curve. |
How similarity is measured
Past cases are compared by which shillers are in, the minute each one entered and the number of mentions.
Confidence
Confidence is a number from 0 to 100. It comes from rules and statistics over the shiller database, not from a language model.
You choose your own threshold in the Telegram bot and only receive signals above it.
Outcome labels
Every past token gets a label from its on-chain history, checked at 1 hour, 24 hours and 7 days. These labels feed shiller win rates and the similar-case search.
| Label | Meaning |
|---|---|
| WIN | The token grew after the calls. |
| RUG | Liquidity was pulled or the price collapsed. |
| DEAD | Trading faded out. |
| NEUTRAL | None of the above. |
The current version of the rules is v1. Rules v2 is on the roadmap.
Shiller database
Each shiller leaves a trail of tokens they called and how those ended. For every shiller Kairos keeps:
| Metric | Meaning |
|---|---|
| SMARTS | Smart followers: how many notable accounts follow the shiller. |
| CA | Memecoins called. Green under 10, yellow 10–19, red 20 and more. |
| Win rate | Share of labelled calls that ended as win. |
| Rug rate | Share of labelled calls that ended as rug. |
| Median minute | Typical minute after token creation when this shiller posts. |
Rates use all of a shiller's labelled calls. Shillers are ranked by their track record.
Telegram bot
On Pump.fun decisions take minutes, so the main channel is a bot, not a dashboard. The bot is in closed beta.
| Feature | What it does |
|---|---|
| Straight to chat | A push within seconds of the shill being found. |
| Your threshold | Set your own confidence threshold and signal filters. |
| The short why | Ask “Why?” for an AI explanation of a strong signal. The score itself still comes from rules and stats. |
| Less noise | Mute shillers you don't trust. |
Dashboard
The dashboard at xkairos.io shows live signals, shiller history and token pages. No sign-up is needed.
Roadmap
Token history is already being collected. Scoring is in progress. Signal delivery comes next.
- New Pump.fun tokens (PumpPortal)
- CA and shiller lookup via Frontrun
- Trade backfill (Helius)
- Outcome labels v1
- Shiller ranking by track record
- Scoring and signals: rules and stats over the shiller database
- Telegram bot, closed beta
- Web dashboard
- Sign-in and subscription
- AI explanations for strong signals
- Outcome rules v2
Glossary
| Term | Meaning |
|---|---|
| Trench | A new memecoin in its first minutes, while early buyers and callers pile in. |
| Shiller | An account that posts a token's contract address publicly. |
| CA | Contract address of a token. Also the count of memecoins a shiller has called. |
| Call | One public post of a CA by a shiller. |
| Migration | A Pump.fun token leaving the bonding curve for a regular pool. |
| Bonding curve | Pump.fun's built-in pricing before migration. Liquidity there is not a regular pool yet. |
| SMARTS | Smart followers of an account. |
| Signal | The Kairos card for one token: who is in, when, similar cases, confidence and risks. |
FAQ
Does Kairos tell me what to buy?
No. Kairos shows who called a token, when, and how their past calls ended. What you do with that is your decision.
Is the score made by AI?
No. The score comes from rules and statistics over the database. AI only explains strong signals in plain words.
How fast is a signal?
Within minutes of the first post with the contract address. The bot pushes it to your chat within seconds of the shill being found.
Why can a signal be wrong?
Past behaviour of shillers does not guarantee the next outcome. New shillers have short histories, and market conditions change.
How do I join the Telegram beta?
The beta is closed for now. Watch xkairos.io for access.
Risk notice
Kairos is an information tool. It is not financial advice and does not recommend buying or selling any asset. Memecoins are extremely risky and most of them go to zero. Past results do not predict future ones. Examples on xkairos.io marked as demo use invented data.