How the Build
Score works
Mogul Build reads the Solana blockchain, X and Telegram for the last 60 minutes and gives every token one number from 0 to 100: the Build Score. It answers one question: are real people buying, and is it growing? Both come from a published formula, and the same inputs always give the same score.
• Mint or freeze authority still active → capped at 25
• Linked wallets hold > 20% of supply → capped at 39
• Chatter up, buyers flat (divergence) → capped at 59
Each component is computed as a pure deterministic function with zero network calls inside the scoring engine.
Build Score
The Build Score is a single number from 0 to 100 answering whether real people are buying a Solana token, and whether that buyer base is expanding.
Hype is cheap to manufacture with bot farms, automated X accounts, and Telegram spam groups. Actual capital commitments on-chain are costly to fake. Mogul Build separates talk from real economic traction by cross-referencing public social volume with on-chain buyer genesis.
The formula runs inside a pure TypeScript engine. Given the exact same inputs, it yields the exact same score. No hidden subjective adjustments, no manual intervention, and no sponsored overrides.
Bands
Every Build Score falls into one of four distinct developmental bands with Mo the Mogul Mole:
Holding a red “CONDEMNED” sign. “Not up to code.” Heavy linked wallets, active authorities, or artificial volume.
Looking at a crack in the wall. “Cracks showing.” Early or mixed signals, sniper presence, or low buyer dispersion.
Standing between traffic cones. “Work in progress.” Substantial organic buying, dispersed wallet funders, authentic social chatter.
Stamps “APPROVED”. “Approved.” Exceptional organic accumulation, independent buyer genesis, and verified organic community talk.
Real Buyers (40)
Takes every wallet that bought the token in the last 60 minutes from trades and asks four questions. If every answer is no, it is a real buyer:
- Q1 — Is it a bundler? Bought in the same launch block as other wallets. (Solana Tracker bundlers)
- Q2 — Is it a sniper or insider? Bots and early insiders who sell to later buyers. (Solana Tracker risk.snipers / risk.insiders)
- Q3 — Is it the dev, or funded by the dev? The creator buying through additional wallets. (Dev wallet + Grok cluster)
- Q4 — Was it funded by the same wallet as 2 or more other buyers? Excludes recognized exchange wallets (Binance, Coinbase, Kraken, etc.). Identifies multi-wallet sybil rings.
Momentum (30)
Compares the last 15 minutes of real buyers against the average 15-minute pace of the last hour:
If realBuyers60 < 5 or pairAgeMin < 15, m is fixed at 0.25 and word is Flat (“Too new for momentum”). If insidersExiting is true, m is capped at 0.25.
Caps
Hard ceiling overrides that cap the total Build Score when structural risks fire. The lowest active cap wins and becomes the verdict line:
Edge Cases
Sources
Mogul Build aggregates data directly from six primary infrastructure sources:
• DexScreener API: Token pair address, liquidity depth, creation age, and price.
• Solana Tracker Data API: Launch bundles, sniper detection, deployer percentage, and 60-minute trade ledger.
• xAI Grok: Sybil / co-funded wallet clustering and social intelligence used in the real-buyer test.
• Official X API (v2): 24-hour mention counts and 50-post sample user authenticity analysis.
• Telegram Listener: Dedicated Python Telethon worker reading mentions across public crypto channels.
• Mogul Build Scoring Engine: Pure deterministic TypeScript calculation module.
Limits
In accordance with Invariant Rule #3, Mogul Build explicitly states what it cannot see:
- Time horizon: Evaluates transactions within the trailing 60-minute window only.
- Funder tracing depth: Traces depth-1 funding origins. Does not trace multi-hop mixers or off-chain escrow contracts.
- Telegram scope: Limited to monitored public groups. Private group chats, invite-only alpha rooms, and DMs cannot be monitored.
- OTC deals: Off-chain agreements and centralized exchange internal transfers cannot be observed.
Versioning
Current specification: Formula 1.0 (Version 4 · 27 October 2026).
Formula changes are versioned, documented publicly, and published with backward-compatible outcome comparisons. Future additions (Influencer Receipts, Telegram Bot, Chrome Extension) will be deployed as separate satellite tools without altering the core deterministic scoring math.
Disclaimer
Mogul Build provides heuristic probability scores and on-chain telemetry. It does not provide financial advice, trading signals, or guarantees of token safety. We never label tokens as scams or individuals as scammers; we output verifiable numbers and observed on-chain states only.
Social Heat (30)
Reads X and Telegram for the last 60 minutes and gives every token two numbers: Heat and Organic %.
v = weighted mentions ÷ max(baseline hourly mentions, 3).V = 100 × min(1, log2(1+v) ÷ log2(21)). If the token has < 6 hours of history, velocity is capped at 60 with tag “New token, limited baseline”.a = unique authors (X sample + Telegram), max 50.B = 100 × min(1, log10(1+a) ÷ log10(51)). 50 authors = 100; 10 = 61.r = median engagement per mention ÷ max(baseline median, 1).E = 100 × min(1, log2(1+r) ÷ log2(11)).C = 100 × min(1, min(authors on X, authors on Telegram) ÷ 10).Each author starts at weight 1.0 and is multiplied by each matching signal (floor 0.05). Organic % = 100 × average author weight. If zero authors, shows “no mentions”.