@agentutility/mcp-matchpoint

@agentutility/mcp-matchpoint

Dedup and record linkage, per call.

Turn messy company names, people, addresses, emails, and phones into stable match keys so duplicates collide — plus a similarity score to rank candidate pairs. The dedup/entity-resolution primitives a data-enrichment agent needs.

Pricing: pay-per-call in USDC on Base. No subscriptions, no API keys. See per-tool prices below.

Install — Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "agentutility-matchpoint": {
      "command": "npx",
      "args": ["-y", "@agentutility/mcp-matchpoint"],
      "env": { "X402_PRIVATE_KEY": "0xYOUR_PRIVATE_KEY_HEX" }
    }
  }
}

Restart Claude Desktop. 7 tools appear in the tool palette.

Install — Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "agentutility-matchpoint": {
      "command": "npx",
      "args": ["-y", "@agentutility/mcp-matchpoint"],
      "env": { "X402_PRIVATE_KEY": "0x..." }
    }
  }
}

Funding

Send any amount of USDC on Base mainnet to the address derived from your X402_PRIVATE_KEY. The MCP server uses it to pay for tool calls automatically.

USDC on Base contract: 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913

Tools (7)

Tool	Description
address-parse	(0.005 USDC/call) Parses a freeform US street address into components: number, street, unit, city, state, and zip. Deterministic, instant, and best-effort — no geocoding or external lookups. Use it as an address parser, address splitter, or USPS field extraction step.
match-key-address	(0.005 USDC/call) Generates a match key for postal addresses with USPS-style standardization so equivalent addresses collide. Expands street types (St to Street), directionals (N to North), and unit designators (Apt/Ste/# to unit), then builds a stable key so '123 Main St Apt 4' and '123 Main Street #4' collide. Deterministic and instant — no geocoding or external lookups. Use it for postal address dedup and address record linkage.
match-key-company	(0.005 USDC/call) Generates a deterministic similarity key for company names so different spellings of the same business collide. Strips legal suffixes (Inc, LLC, Corp, Ltd, GmbH), punctuation, and abbreviations, then phonetic-encodes tokens so 'IBM Corp' and 'I.B.M. Corporation, Inc.' produce the same key. Deterministic and instant — no external lookups. Use it for business-name dedup, entity resolution, fuzzy company matching, and record linkage when deduping CRM/lead lists or joining company records across sources.
match-key-email	(0.005 USDC/call) Normalizes email addresses into a match key so aliases of the same inbox collide. Collapses Gmail dots and '+tag' subaddressing and lowercases the domain so '[e-mail]' and '[e-mail]' share a key; also reports basic syntax validity and the domain. Deterministic and instant — no SMTP or MX lookups. Use it for email dedup, email normalization, or contact record linkage.
match-key-name	(0.005 USDC/call) Generates a match key for person names so variants of the same person collide. Canonicalizes nicknames (Bob to Robert), folds accents, drops titles/suffixes (Mr, Jr, PhD), and is order-insensitive so 'Smith, Robert' and 'Bob Smith' share a key; returns first/last too. Deterministic and instant — no external lookups. Use it for contact dedup, people record linkage, or fuzzy name matching.
match-key-phone	(0.005 USDC/call) Normalizes phone numbers into a canonical E.164 match key so formatting variants like '[phone]' and '[phone]' collide. Strips formatting to digits and returns E.164, country code, and validity. Deterministic and instant — no carrier lookups. Use it for phone dedup, phone normalization, contact record linkage, or as an E.164 formatter.
match-score	(0.003 USDC/call) Scores how similar two values are on a 0-100 scale for fuzzy matching and dedup confidence. Compares two values of a given type (company, name, address, email, phone, or text), blending match-key equality, Jaro-Winkler, and token overlap, and reports whether they share a match key. The companion to the match-key endpoints: bucket by key, then rank candidates by score. Deterministic and instant — no external lookups. Use it as a fuzzy match score, string similarity, or record-pair scoring API.

How it works

• Agent calls a tool (e.g. address-parse).
• MCP server POSTs to https://x402.agentutility.ai/address-parse.
• The endpoint responds HTTP 402 with payment instructions.
• The MCP server signs an EIP-3009 USDC transfer authorization with X402_PRIVATE_KEY and retries.
• CDP facilitator settles on Base.
• The endpoint returns the actual response.

The agent never sees the payment flow — it just gets the result.

Links

• Cluster overview: https://agentutility.ai/matchpoint/
• All MCP packages: https://mcp.agentutility.ai/
• Source: https://github.com/rooz21/x402/tree/main/packages/mcp-matchpoint

Version: 0.1.6 · License: MIT