MCP server that provides access to Interzoid's data matching, enrichment, and validation APIs. Supports 58 tools for company name matching, person name matching, address matching, email validation, phone validation, and other data quality operations. Supports x402 payment protocol for pay-per-call billing.
This MCP server exposes 9 data matching tools from the Interzoid API with consistent naming, clear descriptions, and properly structured schemas. Strengths: all tools follow verb_noun naming (interzoid_X_match/score), descriptions include pricing transparency ($0.01 via x402), and input schemas are well-formed with required/optional parameter distinctions. Weaknesses: descriptions lack context about when to choose between similar tools (e.g., when to use company_match_advanced vs standard company match), output schemas are not documented (responses are opaque JSON from the API), no error handling guidance for API failures, and no parameter constraints (enums, ranges, patterns). All 9 tools follow a consistent pattern with identical structure, reducing cognitive load for users but also indicating missed opportunities for tool-specific guidance.
Generate a single combined AI-powered similarity key from a full name plus address. Records with the same combined key share both a similar address and a similar individual name. Useful for high-precision household/contact-level deduplication. Cost: $0.01 USDC via x402.
Generate an advanced AI-powered similarity key for US street address matching. Handles unit numbers, directionals, and abbreviations. Cost: $0.01 USDC via x402.
Generate a combined AI-powered similarity key from a company name and address for location-specific company deduplication. Cost: $0.01 USDC via x402.
Generate an advanced AI-powered similarity key for company/organization name matching. Names like 'IBM', 'International Business Machines', 'IBM Corp' produce the same key for deduplication and record linkage. Cost: $0.01 USDC via x402.
Generate an AI-powered similarity key for individual/person name matching. Handles variations like 'Bob Smith', 'Robert Smith', 'Smith, Robert J.' producing the same key. Cost: $0.01 USDC via x402.
Output schemas not documented. Tools return raw API JSON responses without any documented structure for return fields, types, or pagination. LLMs cannot predict response shape for chaining or field extraction.
No error handling guidance. Callables return generic mcp.NewToolResultError() with raw API error messages or network errors. LLMs receive no classification (retryable? user-fixable? fatal?) and no recovery suggestions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 68 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 0 | - | v1 |
Compare two individual/person names and receive a match score from 0-100 indicating similarity. Handles name order, nicknames, and abbreviations. Cost: $0.01 USDC via x402.
Generate an AI-powered similarity key for global/international address matching. Handles international address formats and variations across countries. Cost: $0.01 USDC via x402.
Compare two organization/company names and receive a match score from 0-100 indicating similarity. Useful for determining if two company names refer to the same entity. Cost: $0.01 USDC via x402.
Generate an AI-powered similarity key for product name matching. Handles variations in product names, model numbers, and descriptions. Cost: $0.01 USDC via x402.
Missing tool disambiguation. Nine tools exist for overlapping tasks (e.g., company_match_advanced vs company_match, fullname_match vs fullname_match_score). Descriptions do not explain when to use each variant, LLMs must guess or try both.
Parameter constraints missing. 'algorithm' parameters accept undefined variants (e.g. 'ai-deep') with no enum or validation. 'zip' is optional but no guidance on when to include for precision. No regex patterns, ranges, or format specifications.
Combined-operation tools ('address_and_fullname_match', 'company_and_address_match') violate single-responsibility principle. Naming with 'and' signals multiple concerns; better to split into separate tools or make composition explicit in description.
Descriptions lack usage context. All descriptions state WHAT the tool does (generate similarity key) but not WHEN to call it vs. other tools, WHAT a similarity key is used for, or HOW to interpret the output. LLM decision-making is hindered.