API Reference

v1.0.0

AI-powered startup idea evaluation with ZK-verified scoring. Evaluates ideas across 8 dimensions and produces a cryptographically attested score (0-100) via RISC Zero zkVM and zkVerify.

x402 HTTP payment protocol — pay per call in USDC, no subscription needed

Endpoints

POST/api/evaluate

Evaluate Idea (Public)

Evaluate a startup idea publicly. Idea text stored and indexed.

$0.25
USDC / call

Request Body

{
  "idea": "<string, required — min 10 chars>",
  "context": {
    "targetMarket": "<string, optional>",
    "existingSolutions": "<string, optional>",
    "uniqueInsight": "<string, optional>",
    "businessModel": "<string, optional>",
    "successCriteria": "<string, optional>"
  },
  "category": "<string, optional>"
}
Payment network:base-sepolia·Protocol:x402
POST/api/evaluate-secret

Evaluate Idea (Private)

Evaluate a startup idea privately. Idea text never stored — only hash. ZK proof keeps idea as private witness.

$0.50
USDC / call

Request Body

{
  "idea": "<string, required — min 10 chars>",
  "context": {
    "targetMarket": "<string, optional>",
    "existingSolutions": "<string, optional>",
    "uniqueInsight": "<string, optional>",
    "businessModel": "<string, optional>",
    "successCriteria": "<string, optional>"
  },
  "category": "<string, optional>"
}
Payment network:base-sepolia·Protocol:x402

Response Schema

Both endpoints return the same JSON shape. The only difference: idea_url is a public link for /api/evaluate and null for /api/evaluate-secret.

{
  "id": "<uuid>",
  "idea_hash": "<sha256 hex string>",
  "score": "<integer 0–100>",
  "verdict": "<ship_it | decent | needs_work | roasted>",
  "title": "<string>",
  "category": "<string>",
  "sub_scores": {
    "market_analysis": "<integer 0–10>",
    "market_sizing": "<integer 0–10>",
    "technical_feasibility": "<integer 0–10>",
    "business_model": "<integer 0–10>",
    "mvp_scope": "<integer 0–10>",
    "originality": "<integer 0–10>",
    "defensibility": "<integer 0–10>",
    "problem_solution_fit": "<integer 0–10>"
  },
  "evaluation": {
    "market_analysis": "<string>",
    "market_sizing": "<string>",
    "technical_feasibility": "<string>",
    "business_model": "<string>",
    "mvp_scope": "<string>",
    "verdict_detail": "<string>",
    "tagline": "<string>"
  },
  "proof": {
    "status": "<generating | submitted | aggregated | relayed | recorded | failed>",
    "job_id": "<string>"
  },
  "idea_url": "<https://... for public evaluations — null for secret>",
  "created_at": "<ISO 8601 timestamp>"
}

Verdict values

ship_it

Score ≥ 75

decent

Score 55–74

needs_work

Score 35–54

roasted

Score < 35

Score Dimensions

Every evaluation produces sub-scores across 8 dimensions plus a weighted composite score (0–100).

Market Analysis/10

Problem severity, customer definition, and solution-market fit (15% weight)

Market Sizing/10

TAM/SAM/SOM estimation credibility and venture-scale potential (15% weight)

Technical Feasibility/10

Buildability with current technology and engineering risk (15% weight)

Business Model/10

Revenue model viability, unit economics, pricing sustainability (15% weight)

MVP Scope/10

Focused scope, speed to first revenue, go-to-market clarity (10% weight)

Originality/10

Novel approach, unique insight, differentiation from existing solutions (10% weight)

Defensibility/10

Moats: network effects, switching costs, proprietary data, barriers (10% weight)

Problem-Solution Fit/10

Solution directly addresses stated problem, 10x improvement over alternatives (10% weight)

Composite Score/100

Weighted aggregate of all 8 dimension scores, normalized to 0–100

ZK Verification

A RISC Zero zkVM program that evaluates a startup idea across 8 weighted dimensions and computes a composite score (0–100). The proof cryptographically attests that the scoring computation was performed correctly over the provided inputs, binding the idea hash and final score into a verifiable commitment.

What the proof attests

  • ✓Scoring computation was executed correctly over the declared inputs
  • ✓Composite score is derived from 8 sub-scores using fixed dimension weights
  • ✓Idea content hash matches the evaluated input (private evaluation)
  • ✓All sub-scores are integer values in the range 0–10

What it does NOT prove

  • ✗Quality or accuracy of the underlying AI model's judgment
  • ✗Authenticity or novelty of the idea text
  • ✗Future market outcomes or business success
Prover: RISC Zero zkVMVerifier: zkVerifyNetwork: eip155:84532:0x75a7f712635D7918563659795450ddE6751D71BC