Compare
Compare Screwtape.ai
Screwtape.ai compared with other public judge and evaluator services. as of Oct 2026. A cell says not stated when the linked page does not say. This page records public prices and features.
| Service | Price per call | Free tier | Signup needed | x402 networks | MCP | False-consensus detection | Pass/retry/fail verdict | Signed receipts |
|---|---|---|---|---|---|---|---|---|
| Screwtape.ai | $0.005 USD per call for POST /v1/judge and POST /v1/consensus. | 100 calls per UTC day per free API key. FREE_DAILY_LIMIT defaults to 100. | Optional. A Bearer API key or an x402 PAYMENT-SIGNATURE header. | USDC on Base (eip155:8453, 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913) and Solana (solana:5eykt4UsFv8P8NJdTREpY1vzqKqZKvdp, mint EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v). | https://screwtape.ai/mcp | Yes. POST /v1/consensus returns false_consensus_risk of low, medium, or high. | Yes. POST /v1/judge returns pass, fail, or retry. | Ed25519 receipt on a successful judge or consensus call when RECEIPT_SIGNING_KEY is set. Omitted when that secret is absent. |
| crosscheck | $0.02 USDC for a draft up to 12,000 units on POST /v1/check, plus $0.01 per further 12,000 units, up to 48,000 units ($0.05). | No production free tier stated. Base Sepolia test USDC: up to 10 checks per wallet per UTC day. | No account and no API key. | USDC on Base (eip155:8453). Base Sepolia test USDC (eip155:84532). Solana not stated. | https://crosscheckapi.com/mcp | not stated | pass is true or false, plus a list of issues. pass, retry, or fail is not stated. | Ed25519-signed receipt on every paid result. |
| Toll402 judge_output | $0.07 per successful call. | Not for this call. The site states one free test call per IP for a tool priced up to $0.01. judge_output is $0.07. | Not required for x402. A prepaid key (x-toll402-key) is optional. | USDC on Base (eip155:8453). Solana not stated. | https://toll402.dev/mcp | not stated | pass is a boolean. The score runs from 0 to 100, with issues and suggestions. retry is not stated. | not stated |
| Agent402 judge | $0.001 USDC per call, via x402 or MPP. | No free tier for this call, as stated on the judge page. | No signup. Pay per call over x402 or MPP. | The site-wide x402 note lists USDC settlement on Base, Solana, Polygon, Arbitrum, Monad, Celo, Avalanche, Sei, Optimism, Stellar, or Algorand. The judge page does not list a network of its own. | https://agent402.tools/mcp | not stated | Not stated as pass, retry, or fail. Question types are choice, score, and noul (yes/no, probability of yes from 0 to 1). | The site states every settled x402 or MPP response carries a receipt. The judge page does not name a signature algorithm. |
| LangSmith evaluators | Perceived Error Tuned Evaluator: 0.015 LSU per successful evaluation run. The pricing page states 1 LSU = $1, so $0.015. Other evaluator types are not given this flat rate. | Not stated for this evaluator. | Account required. Perceived Error Tuned Evaluator is in public beta on Plus and Cloud Enterprise plans in the US. Plus is $39 per seat per month. | not stated | not stated | not stated | Perceived Error returns true or false. pass, retry, or fail is not stated. | not stated |
Sources
- Screwtape.ai: https://screwtape.ai/llms.txt
- crosscheck: https://crosscheckapi.com/llms.txt
- Toll402 judge_output: https://toll402.dev/tools/judge_output
- Agent402 judge: https://agent402.tools/tools/judge
- LangSmith evaluators: https://www.langchain.com/pricing
- Glama listing for crosscheck: https://glama.ai/mcp/servers/maxugc/crosscheck