LUMORAI Whitepaper
AI-Powered SocialFi Protocol — Full Technical & Economic Documentation
v1.2.0
March 2026
Pre-TGE
BNB Chain
Download PDF
Abstract
LUMORAI introduces a decentralized AI-powered reputation layer for Web3 social ecosystems. By processing multi-platform social signals through a four-layer machine learning pipeline, the protocol produces an immutable, tamper-resistant on-chain reputation score (0–1000) for each participant. This score directly drives LMRAI token emission, governance weight, and protocol access tiers. The system addresses three fundamental failures in existing social infrastructure: (1) value extraction without creator compensation, (2) opaque algorithmic scoring, and (3) absence of portable on-chain social identity.
1. Executive Summary
LUMORAI is a layer-2 social intelligence protocol built on BNB Chain that transforms raw social activity into verifiable, tokenized reputation. The protocol ingests multi-platform data streams from X (Twitter), Telegram, and BNB Chain wallet activity, processing them through a proprietary four-layer AI pipeline to produce a daily reputation score per user.
The LMRAI utility token serves three functions within the ecosystem: (1) daily reward emission proportional to reputation score, (2) governance voting with score-weighted multipliers, and (3) staking to unlock protocol tier benefits. With 15 billion total supply and 40% allocated to community rewards, LUMORAI is structured for long-term participant retention.
Key differentiators versus existing SocialFi protocols include: fully transparent scoring parameters published on-chain, a multi-signal anti-Sybil layer combining biometric heuristics with graph analysis, and a creator economy module (Creator Launchpad) enabling permissionless fan token creation for Tier 3+ users.
0–1000
Reputation Score Range
2. Problem Statement
The current social media landscape exhibits four systemic failures that LUMORAI is designed to resolve:
Value Extraction Without Compensation
Social platforms captured an estimated $150B+ in advertising revenue in 2024, with creators receiving less than 2% through native monetization programs. Influence is commoditized at scale while creators bear the production cost without proportional return.
Opaque & Manipulable Algorithms
Algorithmic reach is determined by proprietary, undisclosed scoring systems that change without notice. Creators optimize for proxies (engagement bait) rather than authentic quality, degrading content ecosystems over time.
Rampant Bot & Sybil Attacks
Industry estimates suggest 15–30% of active social accounts are bot-operated or Sybil identities. Without on-chain verification, reputation metrics are trivially gameable through purchased followers and coordinated fake engagement.
No Portable On-Chain Social Identity
User reputation is siloed within each platform. There is no universal, censorship-resistant social identity layer enabling cross-platform reputation portability or on-chain proof of influence.
Misaligned Web3 Incentive Structures
Existing SocialFi protocols rely on unsustainable token emission models uncorrelated with real user contribution, leading to mercenary farming behaviour and rapid community collapse post-TGE.
3. The LUMORAI Solution
LUMORAI addresses each problem with a corresponding protocol component:
AI Reputation Score → Solves Opacity
A fully transparent, 12-metric scoring model with public parameter weights. Score calculation methodology is published on-chain and auditable by any participant.
Influence-to-Earn → Solves Extraction
LMRAI token rewards distributed daily proportional to verified reputation score. Value flows directly to creators, not intermediaries.
Anti-Sybil Framework → Solves Manipulation
Multi-layer verification combining behavioral biometrics, graph topology analysis, and on-chain transaction history to eliminate fake accounts from score calculations.
On-Chain Identity → Solves Portability
Reputation scores committed to BNB Chain every 24 hours as verifiable on-chain state, enabling cross-protocol reputation imports and censorship-resistant identity.
Score-Weighted DAO → Solves Misalignment
Governance voting power = token balance × reputation score multiplier. Genuine contributors gain disproportionate protocol influence over mercenary actors.
4. System Architecture
LUMORAI operates as a three-tier architecture: Data Ingestion Layer → AI Processing Layer → On-Chain Commitment Layer.
┌─────────────────────────────────────────────────────────┐
│ LUMORAI PROTOCOL STACK v1.2 │
├─────────────────────────────────────────────────────────┤
│ LAYER 3: APPLICATION │ Dashboard │ API │ Launchpad │
│ LAYER 2: ON-CHAIN │ BNB Chain │ Score Commit │ DAO │
│ LAYER 1: AI ENGINE │ NLP │ Graph │ Anti-Sybil │ DeFi│
│ LAYER 0: INGESTION │ X API │ TG API │ BSCScan API │
└─────────────────────────────────────────────────────────┘
Layer 0 — Data Ingestion
Real-time API connectors to X Developer API v2, Telegram Bot API, and BSCScan API. Raw event streams are normalized into a unified social activity schema at ~50K events/minute.
Layer 1 — AI Processing
Four-module ML pipeline (NLP, Graph, Anti-Sybil, DeFi) processes normalized events and outputs weighted score components updated every 15 minutes.
Layer 2 — On-Chain
Daily score aggregations are batch-committed to BNB Chain via a smart contract oracle. Scores are publicly readable and immutable once committed.
Layer 3 — Application
REST API serves score data to the web dashboard, Creator Launchpad, and third-party integrations. All endpoints are rate-limited and authenticated.
5. AI Reputation Engine
The LUMORAI AI engine is a four-module ensemble model. Each module outputs a normalized sub-score (0–100) that is weighted and aggregated into the final reputation score via the formula below:
R =
0.30 × C +
0.25 × E +
0.25 × N +
0.20 × D
R = Reputation Score | C = Content | E = Engagement | N = Network | D = DeFi Activity
30%
Content Intelligence (C)
Transformer-based NLP model fine-tuned on 500K+ Web3 social posts. Evaluates semantic originality, information density, factual accuracy signals, and language quality. Penalizes copied content, low-effort posts, and coordinated narrative spam.
25%
Engagement Intelligence (E)
Measures engagement authenticity via response latency distribution, reply quality depth, and retweet propagation patterns. Bot-generated engagement is identified via temporal clustering analysis and filtered before scoring.
25%
Network Intelligence (N)
PageRank-variant algorithm applied to the follower/interaction graph. Tier-1 follower interactions carry 10× weight vs. anonymous accounts. Network reach score reflects genuine influence propagation, not raw follower count.
On-chain behavioral score derived from transaction consistency, DeFi protocol interactions, token holding duration, and cross-chain bridge activity. Long-term holders score higher than short-term traders.
6. Scoring Model & Metrics
The 12 scoring metrics are grouped by category. All weights are public and governance-adjustable via DAO vote:
| Metric |
Category |
Weight |
Description |
| Content Originality |
Content |
8% |
Semantic uniqueness score vs. corpus of recent posts |
| Post Quality Score |
Content |
7% |
NLP-derived information density and readability index |
| Consistency Streak |
Content |
5% |
Continuous posting cadence bonus (max 30-day streak) |
| Authentic Reply Depth |
Engagement |
8% |
Quality of received replies — filters bot responses |
| Retweet Propagation |
Engagement |
7% |
Network spread velocity of original content |
| Referral Chain Depth |
Engagement |
5% |
Contribution to protocol user acquisition graph |
| Tier-1 Network Reach |
Network |
10% |
Interactions from verified high-reputation accounts |
| Cross-Platform Presence |
Network |
8% |
Multi-platform signal consistency bonus |
| Community Growth Contribution |
Network |
7% |
Net new user additions attributable to account |
| On-chain Transaction Score |
DeFi |
7% |
DeFi interaction breadth and depth on BNB Chain |
| LMRAI Holding Duration |
DeFi |
7% |
Loyalty score based on token holding period |
| Cross-chain Activity |
DeFi |
6% |
LayerZero bridge activity and multi-chain presence |
Score Tiers
LUMEN ELITE: 800–1000LUMEN PRO: 600–799BUILDER: 400–599STARTER: 0–399
7. Anti-Sybil Framework
LUMORAI employs a three-stage Sybil detection pipeline to ensure only genuine human activity contributes to reputation scores:
1
Stage 1 — Behavioral Biometrics
Analyzes posting time patterns, inter-post latency distributions, and vocabulary diversity entropy. Bot accounts exhibit statistically anomalous temporal patterns (Poisson clustering) that are flagged for deeper inspection. Accounts flagged at Stage 1 receive a 0.5× score multiplier pending Stage 2 review.
2
Stage 2 — Graph Topology Analysis
Maps the interaction graph to identify coordinated inauthentic behavior (CIB) clusters. Accounts with abnormally high reciprocal engagement rates (>85% mutual reply networks with <10 unique participants) are classified as potential Sybil rings. Graph analysis runs on a 14-day rolling window using a modified HITS algorithm.
3
Stage 3 — On-Chain Verification
BNB Chain wallet linkage provides ground-truth identity anchoring. Wallets with on-chain transaction age <30 days, zero DeFi interactions, or token purchase patterns consistent with airdrop farming are penalized with a 0.3× score multiplier. Wallets with 6+ months history and diverse DeFi activity receive a 1.2× authenticity bonus.
Appeal Process: Accounts flagged by the Sybil system may submit a manual review request with supporting documentation. The DAO Integrity Committee (5-member multisig) reviews appeals within 72 hours.
8. On-Chain Commitment Layer
Reputation scores are committed to BNB Chain via a batch oracle mechanism at 00:00 UTC daily. The commitment process ensures data integrity and public verifiability:
// LUMORAIOracle.sol — Score Commitment Function
function commitScores(bytes32 merkleRoot, uint256 epoch) external onlyOracle {
require(epoch == lastEpoch + 1, "Invalid epoch");
require(block.timestamp >= lastCommit + 86400, "Too early");
epochRoots[epoch] = merkleRoot;
lastEpoch = epoch;
lastCommit = block.timestamp;
emit ScoresCommitted(epoch, merkleRoot);
}
// Users verify their score via Merkle proof
function verifyScore(address user, uint256 score, bytes32[] proof) external view returns(bool) {
bytes32 leaf = keccak256(abi.encodePacked(user, score));
return MerkleProof.verify(proof, epochRoots[lastEpoch], leaf);
}
Merkle Tree Batch
All user scores per epoch are organized into a Merkle tree. Only the root hash is stored on-chain, minimizing gas cost while preserving full verifiability.
24-Hour Epoch Cycle
Scores are computed off-chain every 15 minutes and finalized at epoch close (00:00 UTC). The oracle multisig (3-of-5) signs and submits the daily root.
Public Verification
Any user can verify their score locally using the published Merkle proof. No trust assumption on LUMORAI is required — pure cryptographic verification.
LMRAI Reward Trigger
The RewardDistributor contract listens to ScoresCommitted events and calculates per-user emission for the epoch, queuing claimable LMRAI balances.
9. Token Economics (LMRAI)
Token Specifications
Token Name
LUMORAI Token
Symbol
LMRAI
Standard
BEP-20
Total Supply
15,000,000,000
Decimals
18
Network
BNB Chain (Chain ID: 56)
TGE Date
Q3 2026 (Est.)
DEX
PancakeSwap
Initial Price
TBD
FDV Target (Est.)
TBD
Allocation
Community Rewards
40%
6,000,000,000
Public Sale
15%
2,250,000,000
Ecosystem Fund
15%
2,250,000,000
Partnerships
15%
2,250,000,000
Team & Advisors
10%
1,500,000,000
Protocol Reserve
5%
750,000,000
Token Utility
Daily Rewards
Primary emission vehicle. Distributed daily to all active participants proportional to reputation score.
Staking
Stake to unlock protocol tiers, boost score weight multipliers (up to 1.5×), and earn base APY (15–40%).
Governance
DAO voting. 1 LMRAI = 1 vote × reputation multiplier. Minimum 10,000 LMRAI to submit proposals.
Creator Launchpad
Tier 3+ users deploy fan tokens. Requires minimum LMRAI stake as protocol fee.
Premium Access
Unlock advanced analytics, priority API, and private signal feeds via LMRAI stake thresholds.
10. Vesting & Distribution Schedule
All non-community allocations are subject to structured vesting to prevent supply shock and align long-term incentives:
| Category |
Allocation |
TGE Unlock |
Cliff |
Vesting |
| Community Rewards |
40% |
5% |
0 months |
48 months |
| Public Sale |
15% |
10% |
0 months |
12 months |
| Ecosystem Fund |
15% |
0% |
3 months |
30 months |
| Partnerships |
15% |
0% |
6 months |
18 months |
| Team & Advisors |
10% |
0% |
12 months |
36 months |
| Protocol Reserve |
5% |
0% |
6 months |
48 months |
11. Ecosystem & Use Cases
Personal analytics hub showing score breakdown, daily delta, tier status, claimable rewards, and historical score trajectory over 90 days.
Tier 3+ users (score 800+) can deploy BEP-20 fan tokens with configurable supply, vesting, and holder utility. Platform handles contract deployment, no coding required.
AI-curated crypto intelligence derived from aggregated community social data. Sentiment shifts, narrative momentum, and wallet movement alerts — available via API to Tier 2+.
Third-party protocols can import LUMORAI reputation scores via the public on-chain oracle to gate access, weight rewards, or adjust collateral ratios based on user trust score.
Native iOS/Android application with push notifications for score updates, reward claims, and governance votes. Social sharing tools for reputation milestones.
RESTful API and GraphQL endpoint for third-party integrations. Rate-limited at 100 req/min (free tier) and 10K req/min (Enterprise). SDK available in TypeScript and Python.
12. Governance (DAO)
LUMORAI transitions to full DAO governance in Q1 2027. The governance model uses a reputation-weighted voting system to ensure genuine contributors have disproportionate influence over protocol direction.
Voting Power Formula
VP = LMRAI_staked × (1 + score/1000). Maximum 2× multiplier at score 1000.
Proposal Threshold
Minimum 10,000 LMRAI staked for 30+ days to submit governance proposals.
Voting Period
5-day voting window. Quorum: 5% of circulating supply must participate.
Timelock
All approved governance changes subject to 48-hour timelock before execution.
Governable Parameters: Metric weights (12 scoring parameters), tier thresholds, daily emission rate, staking APY curves, Creator Launchpad fees, and partnership whitelisting. Core smart contract logic requires a 75% supermajority to modify.
13. Security Model
CertiK Smart Contract Audit
Full audit of all deployed contracts (Oracle, RewardDistributor, Staking, FanToken Factory) by CertiK scheduled for Q2 2026 — 6 weeks before TGE. Report will be published on BSCScan and the LUMORAI website.
Oracle Multisig (3-of-5)
The daily score commitment oracle is controlled by a 3-of-5 multisig. Signers include 2 core team members, 1 independent auditor, and 2 community-elected DAO delegates. Private keys are stored in hardware security modules (HSM).
AI Model Integrity
The AI scoring model weights are hashed and published on-chain at each weekly model update. Any change to scoring parameters requires DAO approval and a 48-hour timelock before deployment.
Bug Bounty Program
Immunefi-hosted bug bounty program launching Q1 2026. Critical vulnerabilities: up to $50,000 LMRAI. High severity: up to $15,000 LMRAI. Scope covers smart contracts, oracle infrastructure, and API endpoints.
Emergency Pause Mechanism
All reward distribution contracts include an emergency pause function controllable by 4-of-5 guardian multisig (separate from governance multisig). Maximum pause duration: 72 hours without DAO vote.
14. Roadmap
Phase 1 — Foundation
Q1–Q2 2026
IN PROGRESS
Protocol architecture & smart contract design
AI engine v1.0 development & internal testing
Community building: 7K+ Twitter + Telegram
Strategic partnership negotiations (8 confirmed)
CertiK audit commission (Q2 2026)
Whitepaper v1.2 publication
Phase 2 — TGE & Launch
Q3 2026
UPCOMING
TGE on BNB Chain via PancakeSwap
Platform beta launch with dashboard & scoring
Influence-to-Earn live for all participants
Oracle deployment & first on-chain score commit
Staking vaults live
Creator Launchpad v1.0 (invite-only)
Phase 3 — Scale
Q4 2026
UPCOMING
Creator Launchpad public launch
Cross-chain expansion via LayerZero
Nansen & The Graph data integration
Mobile app beta (iOS + Android)
CEX listing discussions
Social Signal Feed API (public)
Phase 4 — DAO
2027+
UPCOMING
Full DAO governance activation
Reputation NFT standard (ERC-3525)
Multi-chain reputation protocol (ETH, Solana)
AI model open-source components release
Ecosystem grants program ($5M LMRAI)
1M+ active users target
15. Core Team
EV
Elara Voss
CEO & Co-Founder
Former AI researcher at Google DeepMind. 8+ years in machine learning, social graph analysis, and NLP systems. Led development of two enterprise ML products acquired by Fortune 500 companies.
DH
Dex Hanamoto
CTO & Co-Founder
Ex-Binance Smart Chain core developer (2020–2023). Expert in BEP-20 protocol design, DeFi architecture, and oracle systems. Contributed to 3 top-50 DeFi protocols by TVL.
SY
Sira Yun
CMO
Web3 growth marketer with proven track record at LayerZero Labs and Galxe. Built and managed communities of 200K+ across Web3 projects. Specializes in SocialFi go-to-market strategy.
OS
Oryn Steele
Head of AI
PhD in Natural Language Processing from MIT. 5 years specializing in social signal processing, reputation modeling, and anti-manipulation systems. Author of 3 peer-reviewed papers on social graph analysis.
Partners & Integrations
Aptos FoundationThe GraphLens ProtocolNansenGalxeLayerZeroArkham Intelligence1inch NetworkPancakeSwapCertiKMetaMaskWalletConnect
16. Risk Factors
Prospective participants should carefully consider the following risks before engaging with the LUMORAI protocol:
Regulatory Risk
The regulatory landscape for utility tokens and DeFi protocols remains uncertain across jurisdictions. Regulatory changes could materially impact protocol operations, token transferability, or exchange listings.
Smart Contract Risk
Despite CertiK auditing, smart contracts may contain undiscovered vulnerabilities. Exploits could result in loss of staked funds or disrupted reward distribution.
AI Model Accuracy
The reputation scoring model is subject to adversarial manipulation as attackers discover new evasion techniques. Continuous model updates are required to maintain scoring integrity.
Token Price Volatility
LMRAI token price is subject to extreme volatility common in cryptocurrency markets. There is no guarantee that the token will maintain any particular value post-TGE.
Platform Dependency
The protocol relies on third-party API access (X, Telegram, BSCScan). Changes to API terms, rate limits, or access restrictions could degrade scoring accuracy or protocol functionality.
Adoption Risk
Protocol utility depends on achieving sufficient user adoption. Failure to reach critical mass of active participants could reduce reward sustainability and ecosystem value.
17. Legal Disclaimer
This whitepaper is provided for informational purposes only and does not constitute financial, investment, legal, or tax advice. LMRAI tokens are utility tokens that provide access to specific functions within the LUMORAI protocol ecosystem. They do not represent equity, debt, dividends, profit participation, or ownership interest in any legal entity.
The information contained herein is subject to change without notice. Forward-looking statements are based on current plans and expectations and involve risks and uncertainties. LUMORAI makes no representation as to the accuracy, completeness, or suitability of this information for any purpose.
Participation in any LUMORAI token sale or protocol activity may be restricted in certain jurisdictions. It is the responsibility of each participant to ensure compliance with applicable laws. Please review our full Disclaimer, Terms of Service, and Privacy Policy before participating.