Pay-per-query Claude AI via x402 micropayments. Ask Claude Haiku/Sonnet/Opus, summarize text, review code, translate, analyze sentiment, and get crypto market analysis — all paid with USDC on Base.
Server has 6 tools with complete input schemas and descriptions. Naming follows verb_noun pattern (ask_claude, summarize, code_review, translate, sentiment, crypto_analysis). Descriptions are present and moderately detailed (100-200 chars), meeting baseline minimums. However, parameter descriptions lack depth on constraints and error recovery. Output schemas are not formally documented, responses are inferred from code. Error handling exists but lacks actionable recovery guidance per pattern:recovery-guide. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all tools being READ_ONLY. Pricing information is embedded in descriptions rather than structured metadata.
Ask a question to Claude AI via the AskClaude x402 service. Choose a model: haiku ($0.01, fast), sonnet ($0.03, smart), or opus ($0.10, genius). Payments are in USDC on Base network via the x402 protocol.
Review code and get improvement suggestions using Claude Sonnet via AskClaude. Returns a quality score (1-10), issues with severity levels, and suggestions. Cost: $0.05 USDC per request.
Get crypto market analysis and insights using Claude Sonnet via AskClaude. Analyze tokens, market trends, and trading signals. Cost: $0.03 USDC per request.
Analyze sentiment of text using Claude Haiku via AskClaude. Returns sentiment label (positive/negative/neutral), confidence score (0-1), and detailed analysis. Cost: $0.01 USDC per request.
Summarize text (articles, documents, whitepapers) using Claude Haiku via AskClaude. Cost: $0.02 USDC per request.
Output schemas not formally documented. Code shows responses include 'answer', 'usage', 'summary', 'word_count' fields, but LLMs cannot infer structure without explicit schema documentation. Violates pattern:tool requirement for documented return types.
No tool annotations despite all tools being READ_ONLY. Zod schema definitions lack readOnlyHint metadata. LLMs cannot distinguish safe tools from destructive ones without explicit annotations per current spec (2026-07-28).
Error handling returns generic error messages without recovery guidance. HTTP 402 payment errors include pricing info but lack actionable next steps (e.g., 'Set up x402 wallet on Base network'). Violates pattern:recovery-guide.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | 2026-07-28+ | v2 |
Translate text between any languages using Claude Haiku via AskClaude. Cost: $0.02 USDC per request.
Parameter descriptions lack constraint details. 'target_language' in translate accepts any string but should document format (e.g., 'ISO 639-1 code or full name like Spanish, French, Japanese'). 'tokens' array in crypto_analysis lacks min/max length guidance.
Pricing embedded in descriptions and tool names rather than structured metadata. LLMs may hallucinate costs or fail to track cumulative expenses across multiple tool calls. No cost tracking or budget enforcement mechanism.