Scoring Rubric

The scoring rubric evaluates Integration tier applications across five weighted dimensions. A total score of 70 or above (out of 100) triggers auto-approval with a 48-hour human veto window. Below 70 is auto-rejected with a full score breakdown. Self-score before applying.

The five scoring dimensions

1. Mainnet integration depth 30 pts

How deeply does the project engage with Numbers mainnet primitives (Nid, C2PA, x402, escrow)?

Project calls Numbers mainnet directly (registers assets, issues licenses, queries Nids, anchors C2PA manifests). Multiple primitives combined (Nid + C2PA + x402 or escrow).

Uses a third-party wrapper that obscures the Numbers interaction, or the integration is peripheral rather than core.

2. Milestone verifiability 25 pts

What percentage of milestones have on-chain or deterministic verification?

>75% of milestones are deterministically verifiable; all verification targets are specific and actionable.

<25% of milestones are deterministically verifiable; success criteria are vague or compound.

3. Ecosystem reusability 20 pts

Does the output benefit other builders or agents?

Ships a core infrastructure piece, open-source SDK, or published integration enabling new integration types for others.

Single-use, closed, or private implementation with no ecosystem benefit beyond the applicant.

4. Budget proportionality 15 pts

Is the requested amount proportional to the stated scope?

Amount is conservative relative to scope, well-justified, and reasonable for stated gas/compute cost.

Amount grossly exceeds plausible cost or exceeds cost by >2× without justification.

5. Prior mainnet activity 10 pts

What is the on-chain history depth of the applicant wallet?

>20 mainnet transactions or a prior completed NIP grant. Demonstrates commitment and capability.

Wallet has exactly 1 mainnet transaction (the minimum threshold to pass eligibility).

Self-scoring guide

Dimension	Weight	Your score
Mainnet integration depth	30%	—
Milestone verifiability	25%	—
Ecosystem reusability	20%	—
Budget proportionality	15%	—
Prior mainnet activity	10%	—
Total	100%	≥70 to auto-approve

Fetch the live machine-readable rubric: GET /v1/grants/rubric

Score breakdown in rejection responses

If your Integration tier application scores below 70, the API response includes a full score breakdown:

{
  "application_id": "uuid",
  "status": "rejected",
  "score": 58,
  "score_breakdown": {
    "mainnet_integration_depth": 40,
    "milestone_verifiability": 60,
    "ecosystem_reusability": 50,
    "budget_proportionality": 75,
    "prior_mainnet_activity": 25
  },
  "rejection_reason": "Score 58/100 is below the auto-approval threshold of 70"
}

Use this breakdown to revise your application. A rejected Integration application counts toward your 2-application lifetime limit — review carefully before submitting.

Notes

The rubric applies only to the Integration tier. Experiment tier uses binary eligibility checks only.
Rubric weights may change between program versions. Always fetch the current rubric at GET /v1/grants/rubric.
The scoring model is fully automated. No human scores applications at any threshold.

Contact: [email protected]