ChronoKey MCP server

ChronoKey

Timestamps, UUIDs, and text-to-structure extraction for AI agents, with x402 USDC settlement.

What This Is

ChronoKey is a hosted MCP server that provides three utility tools AI agents need on nearly every run: reliable time context, unique identifier generation, and deterministic text-to-structure extraction. Payment is handled via the x402 protocol — a machine-native payment standard using the HTTP 402 Payment Required status code. Any MCP-compatible client can connect and call the tools.

Endpoint

https://cloudflare-mcp-worker.dylanrenovos.workers.dev/mcp

• Transport: Streamable HTTP
• Method: POST
• Required header: Accept: application/json, text/event-stream
• Content-Type: application/json

Tools

health (free)

Returns "OK". Use it to verify the server is reachable.

timestamp ($0.001 USDC)

Returns the current time in multiple formats.

Input: { "timezone": "America/New_York" }

timezone is optional (IANA name). Defaults to UTC.

Output:

{ "iso8601": "2026-09-26T15:30:00.000Z", "unixSeconds": 1790427000, "unixMilliseconds": 1790427000000, "timezone": "America/New_York", "humanReadable": "Friday, September 26, 2026 at 11:30:00 AM EDT" }

uuid ($0.001 USDC)

Generates 1-100 UUIDs.

Input: { "count": 5, "version": "v7" }

• count (optional): 1-100, defaults to 1
• version (optional): "v4" (random) or "v7" (time-ordered), defaults to "v4"

Output:

{ "version": "v7", "count": 5, "uuids": [ "018f3c8a-1e2b-7c4d-8f1a-2b3c4d5e6f70", "..." ] }

text.structure ($0.008 USDC)

Extracts structured entities from unstructured text and returns a strict, stable JSON schema.

Input: { "text": "Contact alice@example.com on 2026-09-26 about the $42 invoice. See https://example.com for details. #finance @team" }

Output:

{ "emails": ["alice@example.com"], "urls": ["https://example.com"], "phones": [], "dates": ["2026-09-26"], "numbers": [ { "raw": "42", "value": 42, "type": "integer", "negative": false } ], "ipAddresses": [], "hashes": [], "mentions": ["@team"], "hashtags": ["#finance"], "stats": { "wordCount": 15, "charCount": 118, "lineCount": 1 } }

The output schema is deterministic and stable. Every field is always present — empty arrays when nothing matches. There are no external API calls, no LLM, and no rate limits. Input is capped at 20,000 characters.

Fields extracted:

• emails — deduplicated
• urls — http/https, trailing punctuation stripped
• phones — international and US formats
• dates — ISO 8601, with optional time component
• numbers — with type (integer/decimal), raw value, and sign
• ipAddresses — IPv4
• hashes — classified by length (md5/sha1/sha256)
• mentions — @handles
• hashtags — #tags
• stats — word count, char count, line count

Pricing

Supported Networks

All three settle to: 0xAb59e91c7A4e280914681FA8eA2015f2e1826b4f

How x402 Payment Works

1. Send the request without payment.
2. Receive a 402 with a PAYMENT-REQUIRED header. Decode (Base64) to see the accepts[] array — one entry per supported network.
3. Sign a USDC transferWithAuthorization (EIP-3009). Gasless — the facilitator pays the gas.
4. Retry with the X-PAYMENT header.
5. Receive the tool result once the facilitator verifies and settles.

x402 clients handle steps 2-4 automatically.

Discovery

Shortened. The full README is on GitHub.