Test and extract regex matches from text via mcpzoo-regex (uvx)

Recipe: Regex testing and extraction via mcpzoo-regex

Server: mcpzoo-regex (server name: "regex") v1.27.2 via uvx Transport: stdio, NDJSON framing Launch: uvx mcpzoo-regex

Tools

Handshake

→ {"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"agent","version":"1.0"}}}
← serverInfo: {name:"regex", version:"1.27.2"}

→ {"jsonrpc":"2.0","method":"notifications/initialized"}
→ {"jsonrpc":"2.0","id":2,"method":"tools/list"}
← tools: [regex_extract, regex_test]

Example 1: Extract email addresses from text

→ {"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"regex_extract","arguments":{"text":"Contact us at [email protected] or [email protected] for help. Not an email: @mention","pattern":"[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\\.[A-Za-z]{2,}"}}}
← {"count": 2, "matches": ["[email protected]", "[email protected]"]}

Example 2: Extract capture groups (domains from URLs)

→ {"jsonrpc":"2.0","id":4,"method":"tools/call","params":{"name":"regex_extract","arguments":{"text":"Visit https://tani.ai/registry and http://example.com/page","pattern":"https?://([^/\\s]+)","group":1}}}
← {"count": 2, "matches": ["tani.ai", "example.com"]}

Example 3: Test if a string matches an ISO date format

→ {"jsonrpc":"2.0","id":5,"method":"tools/call","params":{"name":"regex_test","arguments":{"text":"2026-06-13","pattern":"^\\d{4}-(?:0[1-9]|1[0-2])-(?:0[1-9]|[12]\\d|3[01])$"}}}
← {"matched": true, "full_match": "2026-06-13", "start": 0, "end": 10}

Example 4: Non-matching test

→ {"jsonrpc":"2.0","id":6,"method":"tools/call","params":{"name":"regex_test","arguments":{"text":"not-a-date","pattern":"^\\d{4}-(?:0[1-9]|1[0-2])-(?:0[1-9]|[12]\\d|3[01])$"}}}
← "不匹配" (Chinese for "no match")

Gotchas

• NDJSON framing — Python server, send one JSON object per line.
• Descriptions are in Chinese — tool descriptions use Chinese text (e.g. "用正则表达式从文本中提取所有匹配内容"), but the tools work fine with any input language.
• "No match" response is in Chinese — regex_test returns "不匹配" when the pattern doesn't match, rather than a structured {matched: false}.
• Python regex syntax — uses Python's re module, so backslash-heavy patterns need proper JSON escaping (double backslash).
• regex_extract returns structured JSON with count and matches array.
• regex_test returns matched, full_match, start, end on success.

When to use this vs. the LLM

Use regex_test/regex_extract when you need deterministic, exact pattern matching — email validation, log parsing, data extraction. Don't use it for fuzzy matching or semantic understanding where the LLM is better.