v1.0-draft · Specification and reference implementation in active development · read the whitepaper
Supply

One billion. Allocated to the people who build the network.

Fixed total supply of 1,000,000,000 KHORR.

BucketShareTokensWho gets it, and how
Usage airdrop 30% 300M Everyone who used the protocol: agents listed, replicas run, attestations tapped. 12% at token generation, 18% metered through ongoing usage rewards.
Emission & community rewards 33% 330M Funds replica rewards on the halving curve. The launch set consumes roughly 96M over its lifetime, leaving about 234M for set expansion and future community programs.
Initial contributors 20% 200M The team and early advisors. One-year cliff, then three-year linear vesting.
Investors 10% 100M Locked on standard terms. Pricing is set with investors, not before.
Khorr Foundation 5% 50M Operations, audits, and the interim arbitration function.
Community grants 2% 20M Builders, integrations, and venue adapters.

Design targets, subject to final calibration and review. Nothing on this page is an offering.

Utility

Two use cases, no third.

Operators bond. Users burn. Holders earn no passive yield. The token's value derives from structural demand tied to work performed or features consumed.

Bond

Replicas, the mint authority, marketplace participants bond the token to acquire the right to run infrastructure. Bond is locked while operating; slashable for provable misbehavior; returned after the 90-day unbond cooling period on exit.

100k token active replica · 10k server shadow · 1k mobile · 1M mint authority

Burn

Users burn the token to mint non-transferable marketplace credits at a fixed USDC price (before token generation, credits are bought directly in USDC). Credits pay for listing slots, priority boosts, capability schema deployment, dispute escalation, premium discovery queries, and verified-operator badges.

Fixed-USDC pricing · Non-transferable · Non-expiring

The token does not appear in the session-binding handshake, the PaymentCommitment, the mint fee, or the audit log. It is structurally insulated from the agent payment layer.

Reward economics

Three tiers, three reward profiles.

Each tier earns a tier-appropriate share. All halve uniformly every twelve months on the same schedule.

Status, honestly. The active replica role runs today in the Khorr Operator app, on test networks, with no bond and no rewards. Shadow attestors and mobile attestors are future design work and are not built. The rewards on this page exist only as design, not in the software.

Tier 1 · Active replica

Quorum participant

Bond100,000
Token rewardRn / epoch
Fee share20% of mint fee
Vesting25% liquid · 75% to unbond
Max slash100%
Unbond90 days

Dedicated servers signing each snapshot under the 3-of-N quorum. Earn the full per-epoch token reward, of which 25% is liquid as earned and 75% stays locked and slashable until unbond, plus 20% of each USDC registration fee split among the replicas that signed. The fee share steps up 10 points at each halving, to a 50% cap.

Tier 2 · Server shadow

Divergence detector

Bond10,000
Token reward0.10 × Rn
Bounty20% of slashed bonds
Max slash100%
Unbond90 days

Permissionless pool of additional server operators running the same protocol in parallel. Signatures provide independent divergence detection. Catch a forging active replica, earn 20% of the slashed bond.

Tier 3 · Mobile

Tap-to-attest witness

Bond1,000
RewardPoints / tap
Cadence1 / epoch
Max slash100%
Unbond90 days

Anyone with a phone, wallet, and the 1,000-token bond. Open the app every hour or so, tap attest; the app signs an attestation for the current snapshot, rate-limited to one per attestation epoch. Attestations accrue points at first, converting to tokens from the airdrop bucket after token generation. Geographic and jurisdictional diversity at scale.

Emission curve

The halving.

Per-epoch reward halves every twelve months. The schedule asymptotes to zero but never reaches it. Operators rely increasingly on the USDC fee share and on token appreciation as the network grows.

Y0 Y1 Y2 Y3 Y4 Y5 R₀ R₀/2 R₀/4 Per replica · per epoch reward

Rn = R₀ / 2n

where n = floor(attestation_epochs_since_launch / 8760)

Emission accrues per attestation epoch (≈1 hour, the cadence attestors witness the snapshot), distinct from the 7-day registry epoch that governs snapshot content. At launch, with R₀ = 1,000 token per attestation epoch and 8,760 attestation epochs per year, each active replica earns 8.76 million token in year one. Halving every twelve months. Asymptotic. As emissions decay, the USDC fee share, which does not halve, increasingly carries operator economics.

The launch set consumes roughly 96 million tokens over its lifetime, comfortably inside the 330M emission reserve. The halving applies uniformly across all replica tiers. When active rate halves, shadow rate halves with it, mobile rate halves with it. The relative gap between tiers stays constant. The absolute supply curve approaches its asymptote.

12mo Halving period
1B Fixed total supply
20% Mint-fee share → replicas
8,760 Attestation epochs/yr
90d Unbond cooling
Marketplace credits

Credits at fixed stable-value prices.

Marketplace credits are non-transferable and denominated in USDC. Before the token launches, credits are simply bought with USDC. After token generation, operators burn KHORR to mint credits at the same USDC prices, which insulates marketplace pricing from token volatility.

Feature Cost (USDC, per 7-day epoch where recurring)
Listing slot
per agent per 7-day epoch
0.50 USDC
Listing priority boost
per tier per 7-day epoch
10 USDC
Capability schema deployment
one-time
25 USDC
Dispute escalation
refundable if prevailing
10 USDC
Premium discovery query
per query
0.01 USDC
Verified-operator badge
per 7-day epoch
5 USDC

Credit prices are design placeholders pending final calibration. Emissions and marketplace burn are calibrated against each other: emissions cap at BOOTSTRAP_EMISSION_CAP per epoch and automatically halt if burn falls below ten percent of emissions over any rolling ninety-day window. Foundation governance can reinstate by supermajority vote with a remediation plan. The HIP-138 trap is closed at the protocol level.

Protocol revenue

The registration fee.

Listing an agent costs a real fee, paid in USDC, not tokens. It starts at $5 and doubles every 25 registered agents, capped at $80, with every parameter adjustable. 20% of each fee goes to the replicas that signed the registration; the share steps up as emissions halve. The fee is the network's spam defense and its long-run operator income in one.

Set growth

Seven triggers, one rubric.

The active replica set starts at five and grows through Foundation-governed expansion when any of seven concrete conditions fires.

≥ 40%

Hosting concentration

On a single hosting provider

≥ 60%

Geographic concentration

In one legal jurisdiction

< 2

Quorum margin

Spare slots before quorum breaks

≥ 2×

Bonded pool depth

Pool ≥ 2× current active set

> 50%

Marketplace growth

YoY registration volume

12 mo

Mandatory review

Since last expansion

!

Security incident

Divergence or attack attempt

Bond lifecycle

Locked while operating. Cooled ninety days on exit.

Bond is slashable for provable misbehavior. Slashing is up to 100% for double-signing; less for lesser misbehavior. Slashed tokens are burned, not redistributed.

01

Bond

Operator transfers token to the bond vault. Locked while operating.

02

Operate

Sign artifacts. Earn rewards. Subject to slashing for provable misbehavior.

03

Exit request

Operator requests unbond. The 90-day cooling period begins.

04

Cooling

Bond remains slashable. Operator continues to operate honestly until release.

05

Released

After 90 days without slashing, the bond returns to the operator's wallet.

Get started

Read the design papers.

Six progressively narrower passes from comparative research to the consolidated replica system design.