Skip to content
Accord

Mechanize the verdict.

Justice is infrastructure.

Two CPI calls — create_dispute() → get_ruling(). Any Solana program gets trustless dispute resolution. Jurors are secret, drawn by VRF, slashed for incoherence. Capture is structurally impossible.

Read the docs →
  1. ORACLE — June 2025: Is that a suit?
  2. FOUNDERS — June 2016: An attacker drains The DAO.
  3. COURTS — October 2022: Avi Eisenberg drains $114M from Mango.

Two calls. Any program.

arbitrable.rs 
// file a dispute from any program
create_dispute(subaccord, options, evidence_hash, fee) → dispute_id

// jurors drawn by VRF · commit · reveal · slash on incoherence

get_ruling(dispute_id) → winning_option

Dispute resolution is a primitive — not a product you integrate, a call you make.

The full lifecycle — create_subaccord → stake → draw → commit → reveal → appeal → finalize — lives in the docs.

Capture is structurally impossible.

01

Secret. Drawn by VRF. Slashed.

Jurors don't exist until the dispute. Drawn by verifiable random function, secret through commit-reveal, slashed for incoherent votes. No standing committee to bribe. No token stash buys the verdict.

02

Honesty is the profitable strategy.

The Schelling Point for a subjective question is truth. Vote your honest belief and it pays. Lie and you bet against everyone's incentive to be obvious.

03

Two CPI calls. One commit.

Any Solana program. No SDK to learn, no governance token to acquire, no committee to brief. The cost of switching to Accord is roughly zero.

An accord, not a committee.

The failure mode, named

March 2025 — one actor with 25% of UMA tokens flipped a $7M Polymarket contract. Token-weighted voting is plutocracy. Accord removes the attack surface.

Lineage

The math is old. The Solana primitive is new.

  1. 1960 Schelling

    Independent agents converge on the obvious answer without communication.

  2. ~500 BCE Athenian sortition

    Draw the judges by lot. A committee can be captured; a random sample can't be bought in advance.

  3. 2019 Kleros

    A Schelling-point court, 1,000+ disputes on Ethereum. Proof the mechanism works. The precedent, not the template.

  4. now Accord

    The same economics, packaged as a composable Solana primitive.

Truth doesn't need a referee — it needs a mechanism that rewards it.

We claim the mechanism. We don't claim traction — v1 is the build target.

The integrator.

The protocol builder who'd rather replace human judgment with aligned incentives than staff a multisig. The Mechanist — who reads the program before the pitch and adopts a primitive by committing a CPI call, not by signing up.

Behind them, when v1 ships: the Juror (earns by being honest) and the Subaccord Creator (founds a scoped arbitration community).

Honest scope

Accord is pre-launch. The v1 instruction set is the build target, not a shipped product. No dispute counts, no TVL, no "trusted by" logos — because none exist yet. Read the spec. Follow the build.

Who's right?

One email when v1 ships on mainnet. That's the whole list.

Read the docs · Join the Telegram