Live example · Ethereum Extitutional

Your community already knows who to trust.

The primitive

There is no one trustgraph.

Each community, application, or protocol defines its own graph, its own rules, and the result it needs.

Anyone can run the computation. A proof lets everyone else check the result without trusting the machine that produced it.

01 · Source graph

Community vouches

become reputation scores.

02 · Compose with

Peer evaluations

give trusted reviewers more weight.

03 · New trustgraph

Contribution funding

produces a proven split of a shared pool.

Trustgraphs compose: scores from one graph weight relationships in another, producing a new, verifiable result.

Working example

One example: a web of trust.

Ethereum Extitutional asks who its community trusts, then turns the answer into something other apps can use.

Step 1

Vouch

Sign a public, weighted vouch. Update or revoke it at any time.

Step 2

Score

Trust compounds: vouches from trusted people carry more weight.

Step 3

Use

Apps can use verified scores for voting power, incentive distribution, or gated chats.

Use cases

Different graphs. Different questions.

  • 01

    Community reputation

    Find standing that comes from relationships, not token balance or a platform-owned rating.

    Graph data
    Vouches between people
    Proven result
    A score for earned trust
  • 02

    Contribution funding

    Let trusted peer judgment direct a shared pool toward valuable work.

    Graph data
    Claims, peer evaluations, and rater reputation
    Proven result
    A funding allocation
  • 03

    Impact discovery

    Surface credible work from public records without trusting a private ranking service.

    Graph data
    AT Protocol follows, claims, evaluations, and acknowledgements
    Proven result
    Scores for people and work
  • 04

    Signer rotation

    Rotate a Safe’s signers with the graph, not a hand-managed admin list.

    Graph data
    A proven score ranking and a signer threshold
    Proven result
    A Safe multisig's signer set

Verification

Don’t trust the computer. Check the proof.

The rules are public, so anyone can recompute the result. A short zero-knowledge proof shows every committed input was included and the result matches those rules.

Every record in the graph
One short proof
The verifier checks the result
An input dropped, or one invented
No valid proof exists
A result is accepted only when the proof matches the graph and its rules.

Roadmap

More inputs. Less exposure.

Trustgraphs should work wherever useful graph data lives, then reveal only what the result needs.

The path runs from public on-chain attestations, through flexible off-chain sources, to fully private graphs that remain verifiable.

  1. 01

    On-chain EAS

    Current

    Public attestations, with every update committed before the graph is computed.

  2. 02

    Off-chain EAS

    Pilot

    Signed attestations without a transaction per edge, anchored so the prover can’t choose the inputs.

  3. 03

    Nostr

    Pilot

    Prove over signed events from relays, using follows and notes people already publish.

  4. 04

    AT Protocol

    Pilot

    Verify repo history and records before computing over social and impact data.

  5. 05

    Private graphs

    Research

    Keep relationships and scores hidden while proving the result is correct.

Start here

Build the next trustgraph.

Start a community vouching network today. Choose its starting accounts, define what a vouch means, and tune how trust flows.

The proof system is permissionless. The platform stays open to new graph programs as the input layer grows.

Create a vouching networkRead current status

Open source

Open source. Take it apart.

Every contract, circuit and page is in the open. Read how it works, then read the code that does it.

Star on GitHubRead the docs