Running a bio-keyword hunt on one chain for a week

Published 8 Jul 2026 · Updated 15 Sep 2026

Quick answer: Pick one chain, run a bio-keyword watch on its name for seven days, and spend the first two of those days tightening filters rather than reading results. That order matters: an untuned watch on a popular chain name produces enough impostors, parody accounts and unrelated uses of the word that most people conclude the method does not work and quit on day one. Tuned, it produces a small daily list of accounts that declared themselves in your niche, most of which have not posted yet. This page is the operating procedure. If you want to know what the tool is first, that is on the Bio Watch feature page.

Why one chain, and why a week

One chain, because the tuning is chain-specific and does not transfer. The qualifier words that clean up a hunt on a new L1 are different from the ones that work on an established chain, and running three at once means you cannot tell which adjustment helped. You have twenty keyword slots on any paid plan; the constraint is your attention, not the quota.

A week, because launches are bursty. A chain can produce nothing for three days and eleven accounts in an afternoon after an announcement. Judging the method on 24 hours tells you about that day's news cycle, not about the method.

Choose a chain in the awkward middle: too established and you drown in accounts that mention it in passing; too obscure and there is genuinely nothing to find. The sweet spot is a chain people are actively building on but which has not yet become a word everyone puts in their bio.

Day 1 — the deliberately naive run

Start with just the chain name and no filters at all. This feels wrong and it is the right move: you are not hunting yet, you are sampling the noise so you know what you are filtering against. Guessing your filters before seeing the noise means filtering out the wrong things.

Point it at a channel you can ignore. Then read what arrives and sort it into buckets. On almost any chain name you will find:

• Real new projects — the target. Small follower counts, recently created, a bio that states what they are building.
• Personal accounts — developers and traders who list the chain as an affiliation. Not launches, but occasionally worth following.
• Parody and fan accounts — usually announce themselves in the bio, which makes them cheap to exclude.
• Impostors — the interesting category. Accounts imitating something that already exists, which is itself a signal that the thing being imitated is getting attention.
• Coincidental uses — the word means something else. Chains named after ordinary English words generate a lot of this.

Do not act on anything today. You are calibrating.

Day 2 — tighten it

Now use what you saw. Three controls, applied in this order:

1. Qualifiers. Require the bio to also contain one of a set of words — something like nft, dex, meme, token, protocol. This is the highest-leverage single change, because it converts "mentions the chain" into "is building something on the chain", which is the actual question. It maps directly onto how founders write bios: first NFT on <chain>, native DEX for <chain>.
2. Exclusions. Add the stop-words you saw yesterday — parody, fan, not affiliated. Cheap and effective, because parody accounts label themselves.
3. Minimum followers. Apply this last and keep it low, near zero to single digits. This is the control most likely to remove the thing you are hunting for, since a genuinely new project has no followers on its first day. Its real job is removing empty shells, not filtering for quality.

Matching is word-boundary aware, so a keyword like tempo matches @tempo, on tempo and tempo network but never contemporary. That removes an entire class of noise before you configure anything.

Days 3 to 7 — reading the results

By day three the volume should be low enough to read every alert, which is the target state. If you are still skimming, tighten further — a list you skim is a list you will stop opening.

An alert looks like this:

Three fields carry almost all the information, and reading them in this order is faster than reading the bio first:

• The badge. FRESH means a newly-created account — the high-signal case. The alternative badge means an established account that just added your keyword, which is a pivot or a rebrand. Both fire, and they mean different things: a fresh account is a launch, a pivot is a team changing direction.
• Account age against follower count. Two days old with forty followers is a normal new project. Two days old with forty thousand followers is a bought account, and worth treating with suspicion rather than excitement.
• The bio itself, with your keyword highlighted in the account's real words. Specific claims — a named mechanism, a stated function — read differently from generic ones.

What you do with a hit is a separate discipline: check whether the account follows or is followed by anyone credible, whether the name is already taken by something real, and whether the project exists anywhere outside X. Bio Watch tells you an account appeared. It does not tell you the account is honest.

What this method reliably surfaces

• Projects on day zero, before the first post. The account is created and the bio is written hours or days before anything is published — that gap is the entire edge.
• Rebrands and pivots. An established team adding a chain to its bio is a decision that usually precedes the announcement.
• Ecosystem waves. Several accounts appearing in one week around one chain is itself the signal, independent of any individual account. Attention is arriving.
• Relaunches and impostors. A rugged project respawning under a new handle keeps the name in its bio, because that is the point.

What it misses, and you should expect this

The failures are systematic rather than random, which at least makes them plannable:

• Projects that never name the chain in the bio. A bio reading only "building something" matches nothing. Plenty of the best teams are deliberately vague early, and they are structurally invisible to this method.
• Stealth launches. A team that creates the account and writes the bio the same minute as the first post gives you no lead time at all.
• Chains named after common words. The noise floor may simply be too high, and qualifiers can only do so much.
• Anything that starts somewhere else. Projects born in Discord, Telegram or a launchpad and only later given an X account are found late or not at all.
• The keyword you did not think of. If a chain's community uses a nickname or a ticker rather than the formal name, watching the formal name misses the conversation. Run both if the slots allow.

The honest summary: bio hunting produces a steady stream of real early signals with a low hit rate per alert. It is a sourcing method, not a decision method. Anyone promising otherwise is selling something.

Running it alongside the other two methods

Three different signals answer three different questions, and they arrive in this order:

The follow graph is the useful confirmation step for this workflow. A fresh account from a bio hit is a maybe; a fresh account that a serious builder followed within a day is a different proposition. Running the second as a check on the first is more productive than running either alone.

Setting the hunt up

1. Open @x_relay_bot and tap 👀 Bio Watch, or on Discord run /addbio keyword:<chain> webhook:<url>.
2. Run it bare for a day, pointed at a channel you do not mind ignoring.
3. Add qualifiers and exclusions on day two — /addbio keyword:monad webhook:<url> qualifiers:nft,dex,meme.
4. Read every alert from day three onward. If you cannot, tighten rather than skim.
5. Remove a watch with /removebio when the chain goes quiet, and reuse the slot.

Bio Watch is a paid feature and is not part of the free trial. Every paid plan from Lite ($15/mo) upward includes it, with up to 20 bio keywords.

FAQ

Bio Watch feature → · Monitoring X for alpha → · Guides and comparisons → · Pricing →