What this is,
and how to use it

The metrics that exist — impressions, likes, replies, follower count — measure how a post performed as a post. None of them answers the only question the money was spent on: did it bring anyone in? Those two things come apart constantly. A post with 900 replies can move nothing. A post with 40 can move six figures of volume.

OpenKol measures the second thing. Paste a post; get back what happened after it — volume and distinct traders that arrived, how the token did against comparable tokens over the same window, and whether the engagement underneath it was real. Four separate answers, deliberately never averaged into one, because a single number would hide exactly the disagreements that are worth seeing.

Who it is for

Projects deciding who to pay next, and needing the last round to be defensible to whoever signed it off. KOLs whose numbers are real and who currently have no way to prove it that a buyer would believe. Traders working out whether an account they follow calls anything that goes anywhere.

What it does not do

It does not predict. Every number here is a measurement of something that already happened, and a good score is not a signal to buy — the move it describes is over. It does not rank people by how much you should trust them, and it does not decide whether a call was made in good faith. It measures what a post did, and stops there.

And the two things underneath them

Peer-relative scoring. Performance is measured against what similar tokens on the same chain did over the same hours. A token that doubled during a week when everything doubled has not demonstrated anything, and scoring it as though it had is how a measurement system ends up rewarding luck.

Calibration. Scores are placed against the distribution of every call already measured, so a 62 means something relative to the field rather than relative to a scale someone picked. The percentile is the part that needs a paid plan.

1. 01
   Paste a post on the home page

   Any link to a post on X that mentions a token. You will be asked to confirm you are a person — the check spends a shared budget, and that gate is what keeps a script from spending all of it.
2. 02
   Wait for it

   Twenty to sixty seconds, depending on how much history the token has. The page shows progress as it goes; you can leave and come back to the same URL.
3. 03
   Create an account

   Email and a password. Confirming the address is optional and nothing is restricted without it — it exists so a forgotten password is recoverable.
4. 04
   Check from the dashboard

   Signed in, there is a check bar on the overview page and your history underneath it. The allowance follows the account, not the browser, so it survives clearing cookies and moving machines.
5. 05
   Raise the allowance, if you want more

   Either by holding the token — link a wallet and the ladder below applies — or with a plan. The two do not stack; whichever gives you more is the one that counts.

The token

Every report opens with the coin it measured — logo, ticker, name, chain, contract, and a live price and market cap. This is first because the single most damaging thing this product can do is measure the wrong token under someone's name, and a report that said only $BRODIE gave you no way to notice.

Tickers are not unique. When a post names one that several tokens claim, the resolver picks on liquidity, age and the pool that existed when the post went out — and it can still get it wrong. Anyone signed in can correct it by pasting the right contract address. The replacement has to be a token whose symbol the post actually names, which is what stops a report being reattached to something unrelated.

The four answers

Shown side by side and never summed. A post can score high on impact and low on performance — it brought people in and the token went down anyway — and that pair of facts is more useful than any single number that could be made out of them.

What the scoring version is for

Every stored result carries the version of the scorer that produced it. When weights change, old reports are not silently rewritten to match the new ones; they keep saying what they said, and the changelog marks which changes altered stored results. A report you linked to six months ago still means what it meant when you linked it.

What is missing, and said so

When part of the evidence could not be gathered, the report says which part and scores without it rather than substituting a guess. A missing input is visible on the page. It is never quietly replaced by a plausible number.

Holding the token instead

Linking a wallet raises the free allowance in proportion to what it holds, valued in USD. Valued rather than counted, because a fixed number of tokens means something different every week and a ladder written in tokens quietly re-prices itself whenever the market moves.

The rules around it

One wallet counts for one account. Unlinking a wallet does not immediately free it for a different account — otherwise a single funded wallet could be walked down a list of accounts, granting each of them a tier in turn, and the whole ladder would cost one bag. Linking asks for a signature, never a transaction: it costs no gas, moves nothing, and approves nothing. The balance behind the address is then read on our side, because a number your browser reports is a number your browser can invent.

Holding raises the allowance and nothing else. The corpus percentile, track records and exports stay with the paid plans — making them purchasable in a second currency would quietly make the pricing page untrue. And the two do not stack: you get whichever of your plan and your holding is larger.

What it does

Holding it raises your free daily checks, on the ladder published under Plans & limits. That is the whole utility, and it is deliberately the kind that cannot be faked: the allowance is granted by reading the chain, so the thing the token buys is the thing the token is.

What it deliberately does not do

No revenue share, and no yield. Not as a cost-saving measure — a token that pays holders out of company income is a security in most of the places our customers live, and the version of this product that survives contact with a regulator is the one that never pretended otherwise.

No vote over the scoring. The methodology is published and fixed, and holders cannot change a weight. A measurement that the measured can vote on is worth nothing to the person reading it — and the people with the most tokens would be the people with the most to gain from a kinder scorer.

No presale and no private round. There is no allocation sitting below the market waiting to be sold into it.

None of this is investment advice, and the token is not an investment product. It is a key that unlocks a larger free tier. Buy it if you want the allowance; if you want the product, the plans do more and cost less to reason about.

Vocabulary

Check

One post, scored. The unit everything else is built from.

Impact

Volume and distinct traders brought in against the pair's own baseline, out of 40.

Performance

Peak, hold and drawdown in excess of peer tokens on the same chain, out of 60.

Audience

Reply-content classification — real, filler, or farmed, out of 40.

Profile

A label, not a score: engagement and volume per thousand followers against the measured p75 of each.

Percentile

Where a check sits in the corpus distribution. The part that needs a paid plan.

Intent

Whether a post is a call at all. About 31% of token posts are not.

Scoring version

Stamped on every stored result, so an old report stays reproducible.