HivemindOS manual
KOL Radar
KOL Radar watches the crypto accounts that move markets and tells you, fast, when one of them touches your project. Open More → KOL Radar to see engagements as they happen, the price impact measured afterwards, and the web of connections between the accounts you track.
Watch a Project
Add your project’s X account and its token contract (Base, Solana, Robinhood Chain, or Ethereum). KOL Radar resolves the deepest liquidity pool and starts recording the price, liquidity, volume, and buy/sell counts every scan, so every engagement can be compared against a real baseline.
Extra keywords (a ticker nickname, a product name) widen what counts as “talking about the project”. Telegram groups tied to the project can be added here too.
Track Accounts
Add the X accounts you have seen move your token. Each account is scored on reach, engagement rate, amplification, posting cadence, and verification, and given a tier (S, A, B, C). Its chain focus (Base, Solana, Robinhood Chain, Ethereum, or multi-chain) is read from its bio, links, and last hundred posts, with the evidence shown next to the label.
Accounts that engage a watched project and clear the follower floor you set are added on their own, so the map grows from your project outward.
How Follows Are Detected, and What It Costs
X bills each user object it returns, so KOL Radar lets you choose how new follows are noticed, and shows the estimated daily cost next to each option:
- Newest followers of each project (default) reads a handful of your project’s most recent followers each check. Repeated reads of the same followers are free within a day, so this costs a few cents a day plus a cent for each genuinely new follower. It catches anyone following your project, tracked or not, but cannot see unfollows.
- Every tracked account’s following list snapshots who each tracked account follows. It also sees unfollows and feeds the connection web’s “follows” links. This read can run on either of two rails, and KOL Radar shows the price of each:
- Apify bulk rail scrapes the lists through a pinned Apify Actor for a fraction of a cent per follow. It is the affordable way to keep the full follows graph, and it needs an Apify API token. The data is a few minutes old, and it comes from scraping rather than the official API.
- Official X API is fresh and official, but each of the thousand users on a page is a billed resource, which makes a full graph refresh expensive.
- Off skips follow detection; posts, replies, reposts, tags, and Telegram are still tracked.
Separately, a daily X spend ceiling stops every X read once the day’s estimated spend reaches the amount you set, and the scanner shows today’s running estimate.
What Counts as an Engagement
A follow (or unfollow) of the project account, a reply, a quote post, a repost, a tag, a post naming the token address, ticker, or a keyword, and a Telegram message in a tracked group that names the project. Each one becomes an event with the account’s reach at that moment.
If something happened before KOL Radar was watching, use Record a past engagement and the impact is filled in from minute-by-minute price history around that time.
Data Rails and Redundancy
KOL Radar reads from three sources, and each kind of lookup goes to the cheapest one that can serve it, falling through to the next when one fails or runs out of credit. The Data rails card shows which are connected, what each can do, and how much each lookup costs.
- The official X API is the only source that can read timelines and mentions as they happen, so it drives the live watch loop.
- Apify is the only affordable way to read full following lists, and it is the fallback for profile lookups.
- ScrapeCreators is the cheapest way to look up a profile by handle, and repeated lookups of the same account cost nothing for a week.
Where two or three rails overlap, any one of them can fail without stopping the feature. Where only one rail can do the job, the card says so rather than implying a fallback exists.
If a rail fails repeatedly, empties its credit, or runs low, you get a message on Telegram naming the rail, what broke, which rail is covering for it, and how to restore it. The same warning appears in Notifications. A failing rail is set aside for fifteen minutes and then retried automatically.
Alerts and Impact Reports
Two messages follow every fresh engagement:
- The fast alert, the moment it is seen, with the account’s tier and reach, what it did, and the token’s price, liquidity, and flow right then.
- The impact report once an hour has passed (and a final one at 24 hours), with the price change at 5 minutes, 15 minutes, 1 hour, 4 hours, and 24 hours, a verdict (pump, lift, flat, or dump), and buy/sell flow.
Alerts use the same messaging channels as other HivemindOS escalations (Telegram, Discord, Slack, and the rest) and always land in Notifications. Catch-up scans of old posts never send alerts.
Connection Web and Sequences
The Connection web draws accounts and projects as a map. Arrows run from the account that moved first to the ones that followed inside the sequence window, dotted lines join accounts that engaged the same project together or share a Telegram group, and faint lines show who follows whom. Click any node to see its links.
Shared follows (in the Connection web tab) lists the accounts that three or more of your tracked accounts follow, built from following lists already read. Those are the hubs and cohort-mates worth tracking next. Names for unresolved ids can be looked up on demand, and the button shows the price before you press it.
The Sequences tab lists every case where two or more tracked accounts engaged the same project inside the window, with the lead account, the gaps between steps, and the best measured impact. Each account’s table row shows how often it led versus followed.
Telegram
Telegram only lets your own account read the groups you belong to, so KOL Radar’s group listener signs in as you, read-only, and streams new messages from the groups tied to your tracked accounts and projects. The Telegram tab shows the listener status, which groups are watched, the exact commands to set it up, and the matched messages.
Settings
Under Scan and alert settings you can change how often scans run, the sequence window, the follower floor for automatic discovery, and whether fast alerts and impact reports are sent.