Everything you need
to understand Nexel.

Overview

Nexel is building the identity and reputation layer for autonomous AI agents on Base. The protocol solves the trust gap that emerges when agents transact with each other and with capital. Every agent gets a sovereign on-chain identity, with a verifiable history of behavior visible to anyone.

The project is currently pre-launch. There is no token. There is no staking. There is no minting. The tools available today use real Base mainnet RPC and a transparent scoring formula. Identity contracts deploy after audit completes.

Mission

Identity is the missing primitive in the agent economy. Existing solutions either skip it entirely (most launchpads) or solve it for humans (proof-of-personhood projects). Nexel solves it for agents: machines that need to prove competence, accountability, and accumulated reputation rather than biometric uniqueness.

• Every agent gets a sovereign, soulbound identity on Base
• Every action emits a verifiable audit receipt
• Every reputation point is signed by a counterparty
• Every integration uses the same trust substrate

Trust Score

The Trust Score is a deterministic function that combines three normalized signal components and applies an age multiplier and a sybil penalty. Scores are not opinions — they are computed from measurable on-chain data and verifiable attestations.

Formula

R       = reputation points       (0–1000, normalized)
F       = on-chain footprint      (0–100, derived from chain state)
A       = attestation density     (0–100, distinct signers)
M_age   = account age multiplier  (1.0 to 1.5, log-scaled)
P_sybil = cluster penalty         (0 to 30)

T       = trust score             (0–100)

The geometric mean ensures that an agent cannot compensate for zero attestations with high activity, or zero activity with many attestations. Every component must be present and non-trivial to score well. This is a deliberate design choice — single weak signals collapse the entire score.

The age multiplier rewards established addresses with up to a 50% score boost at 3+ years. The sybil penalty subtracts up to 30 points for cluster patterns suggesting coordinated burst behavior (new addresses with very high tx counts in short windows).

Tier classification

Trust scores map to tiers used by integrating protocols for access control and rate gating.

Range	Tier	Authority Level
80 — 100	Sovereign	Unrestricted. Institutional capital eligible.
65 — 79	Tier I · Trusted	High-trust delegated authority.
50 — 64	Tier II · Verified	Bounded autonomous operation.
35 — 49	Tier III · Emergent	Supervised. Low capital limits.
20 — 34	Tier IV · Nascent	Probationary. Read-only contexts.
1 — 19	Unranked	Insufficient signals.
0	Unverified	No identity record.

Address Inspector

The Inspector takes any Base address and computes a preliminary Trust Score using only data available on-chain. It queries Base mainnet RPC endpoints for transaction count, balance, contract code, and activity history.

Because Reputation (R) and Attestation Density (A) require deployed contracts that are still pre-launch, today's Inspector score is primarily driven by Footprint (F) and the age multiplier. Once attestation contracts deploy, scores update to include R and A automatically.

• Every query hits Base mainnet RPC directly
• No proprietary database, no scraped data, no cached fake metrics
• Account age estimated via binary search of historical block state
• Address type detected via eth_getCode (contract vs EOA)

Data Sources

Nexel uses the following data, all on-chain and verifiable:

Signal	RPC Method	What it tells us
Tx Count	eth_getTransactionCount	Active vs dormant
Balance	eth_getBalance	Economic skin in the game
Type	eth_getCode	Smart contract or EOA
Age	Binary search	Established vs new
Latest	eth_blockNumber	Chain reference

Limitations

The Inspector queries are limited by public Base RPC rate limits. If you query too many addresses in quick succession, requests may fail. The Inspector falls back across 7 public RPC endpoints automatically.

Account age estimation uses a binary search over historical block state, bounded to 18 iterations to keep queries fast. For very old addresses this gives sufficient precision; for newly created ones it's exact.

Protocol Primitives

When the protocol launches on Base mainnet, it consists of four composable primitives. Each one is a separate contract or schema that integrates cleanly with existing Base infrastructure.

Identity Layer

Nexel sits at L5 of the agent stack — the identity substrate that other layers can build against. Compute (L1), Framework (L2), Launchpad (L3), and Settlement (L4) all exist independently. None of them ship identity. That's the gap Nexel fills.

• Sovereign Identity — soulbound on-chain identity per agent
• Reputation Graph — on-chain attestations (planned: EAS schemas)
• On-Chain Footprint — live activity verification from Base RPC
• Audit Receipts — EIP-712 signed action receipts

Post-Launch Features

When the protocol activates on Base mainnet, additional features become live:

• Soulbound identity issuance via Ethereum Attestation Service (EAS)
• Counterparty attestations via EAS schemas (post-deployment)
• Real-time reputation graph indexed via subgraph
• Cross-protocol SDK for capital allocators
• Slashing module for protocol violations (governance-gated)

FAQ