Are you an LLM? Read llms.txt for a summary of the docs, or llms-full.txt for the full context.
Skip to content

On-Chain Reputation

Deterministic trust scores derived from channel history.


AMP includes a reputation system derived from channel history. Scores are computed on-chain from verifiable settlement data — no external oracle, no subjective rating.

Score calculation

base_score = (channels_completed / channels_opened) * 5000
volume_bonus = min(log2(total_volume / 1_000_000) * 500, 2500)
streak_bonus = min(current_streak * 50, 1500)
dispute_penalty = dispute_count * 500
 
score = clamp(base_score + volume_bonus + streak_bonus - dispute_penalty, 0, 10000)

Score range: 0 to 10,000. Higher is better.

ReputationAccount

FieldTypeDescription
entityPubkeyAgent or service being scored
total_channels_openedu64Lifetime channels
total_channels_completedu64Clean closes
total_volumeu64Lifetime settlement volume
total_settlementsu64Successful settlements
dispute_countu64Disputed closes
current_streaku64Consecutive clean settlements
scoreu64Computed score (0-10000)

Reputation-based pricing

Services can offer tiered pricing based on agent reputation:

{
  "pricing": {
    "default": { "mode": "per-call", "rate": "10000" },
    "reputation_tiers": [
      { "min_score": 5000, "rate": "8000" },
      { "min_score": 8000, "rate": "5000" },
      { "min_score": 9500, "rate": "3000" }
    ]
  }
}

Higher reputation = lower prices. Incentivizes good behavior.

How scores update

EventEffect
open_channelIncrement total_channels_opened
settleIncrement total_settlements, extend streak
close_channel (clean)Increment total_channels_completed
close_channel (dispute)Increment dispute_count, reset streak