agentreferee.ai

https://agentreferee.ai/review-3x3

Shares a payout wallet with 22 other endpoints — its adoption is attributed from that shared wallet's traffic, not earned exclusively.

Agent Referee multi-part sealed review — one piece of work in 3 pieces (up to 12000 characters each, 36000 in all): three independent AI reviewers from three different companies each read every piece in one call, in parallel, without seeing each other's verdicts; majority verdict, every vote in a signed receipt with each piece's hash. One agent, one call; charged only if the review completes.

26
reliability

Reliability & health

Uptime and response time from our own unpaid probes, plus the reliability score built from them.

Uptime

0.0% avg
0% 50% 100% Oct 29Nov 5Nov 10Nov 17

Response time

0ms 50ms 100ms Oct 29Nov 5Nov 10Nov 17
Data table
Reliability history per time bucket over the last 30 days (daily).
Bucket Uptime Probes Response time
10/29/2026, 12:00:00 AM 0% 1 —
10/31/2026, 12:00:00 AM 0% 1 —
11/2/2026, 12:00:00 AM 0% 1 —
11/5/2026, 12:00:00 AM 0% 1 —
11/7/2026, 12:00:00 AM 0% 1 —
11/10/2026, 12:00:00 AM 0% 1 —
11/12/2026, 12:00:00 AM 0% 1 —
11/15/2026, 12:00:00 AM 0% 1 —
11/17/2026, 12:00:00 AM 0% 1 —
26
Composite reliability — the weighted average of the dimensions below, on a 0–100 scale.
Availability
0

How often it was reachable when we probed it (weight 35%).

Latency
not yet measured

How quickly it answers when it is up — fast scores high (weight 20%).

Security
100

Whether it serves over working TLS / https (weight 15%).

Transparency
55

Whether the operator disclosed what it does — description, schema, identity (weight 15%).

Activity
4

Observed on-chain payments to this endpoint's wallet — transactions, volume, and distinct paying wallets over the last month (weight 15%; not yet measured when we have not seen its wallet settle).

Confidence 18% — how much probe evidence backs this score. A thin history is flagged, not hidden.

See the scoring methodology for exactly how each number is computed.

Payment

Currencies
USDC
Networks
base
Facilitator
unknown
SDK
unknown

Operator

Company
inferred — unknown
Jurisdiction
—
Registrar
—
Domain registered
—
Hosting network
—
Hosting country
—

Operator context is derived from public RDAP/TLS records and unverified. Company and country are only available for endpoints that publish them (an OV/EV TLS certificate or a non-redacted domain record); most endpoints are CDN-fronted and disclose neither.

On-chain activity (base)

1,344
payments (30d)
145.92
USD volume (30d)
1
paying wallets (30d)

Payments per day

last payment 11/1/2026
0 200 400 Oct 22Oct 25Oct 29Nov 1
Data table
Settled payments per day.
Day Payments Volume (USD) Paying wallets
10/22/2026 8 0.28 1
10/23/2026 324 3.25 1
10/24/2026 106 1.34 1
10/25/2026 146 1.84 1
10/26/2026 155 12.53 1
10/27/2026 82 11.95 1
10/28/2026 334 42.53 1
10/29/2026 113 35.1 1
10/30/2026 68 31.9 1
10/31/2026 2 1.4 1
11/1/2026 6 3.8 1

This payout wallet is shared: figures cover all 23 endpoints paid at it, since payments cannot be attributed to one of them. Only payments settled on chains and facilitators we index.

Response schema
{
  "input": {
    "bodyFields": {
      "criteria": {
        "description": "Optional weighted criteria, e.g. [{\"name\":\"accuracy\",\"weight\":3}]",
        "required": false,
        "type": "array"
      },
      "pieces": {
        "description": "Exactly 3 items {name?, text}: text 1-12000 chars, 36000 in all; name optional, 1-80 chars",
        "required": true,
        "type": "array"
      },
      "task": {
        "description": "What the agent was asked to do (10-4000 chars)",
        "required": true,
        "type": "string"
      },
      "title": {
        "description": "Optional short title",
        "required": false,
        "type": "string"
      }
    },
    "bodyType": "json",
    "discoverable": true,
    "method": "POST",
    "type": "http"
  },
  "output": {
    "receipt": {
      "description": "Ed25519-signed receipt (receipt_format review-multi-v1) with hashes of the task and each piece",
      "type": "object"
    },
    "result": {
      "description": "verdict: pass | needs_changes | fail | cannot_assess | split, plus the tally",
      "type": "object"
    },
    "votes": {
      "description": "Each reviewer's verdict, issues found (each starting [piece N] or [all]), reasoning and scores",
      "type": "array"
    }
  }
}