Product Updates

ApiCrate v0.4: What’s New

ApiCrate Team
api mcp

ApiCrate v0.4 is our biggest release yet. Here's what's new since v0.2 — new APIs, faster parsing, spatial search, and a much better MCP experience.

Email Risk Scoring

New in v0.3, now refined in v0.4. Check any email address for fraud signals before it hits your database:

curl -X POST https://api.apicrate.io/api/v1/email/risk \
  -H "X-API-Key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"email": "user@tempmail.org"}'

The API checks for disposable domains, validates MX records, and returns a risk score. It also normalizes email addresses with provider-aware canonical forms — so j.doe+spam@gmail.com and jdoe@gmail.com resolve to the same identity.

Available as both REST endpoints and MCP tools (apicrate-check-email-risk and apicrate-check-email-risk-bulk).

The postal code API now supports geographic queries. Find postal codes near any point on earth:

curl -X POST https://api.apicrate.io/api/v1/postal/GB/nearby \
  -H "X-API-Key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"lat": 51.5074, "lng": -0.1278, "radius_km": 5}'

New endpoints and MCP tools in this release:

  • Find nearby postal codes — POST /api/v1/postal/{country}/nearby — search by coordinates and radius
  • List postal systems — GET /api/v1/postal — browse all 250+ supported countries
  • Get postal system details — GET /api/v1/postal/{country} — format rules, validation patterns, and metadata per country
  • Bulk validation — POST /api/v1/postal/validate — validate up to 100 postal codes in a single request

All postal data is spatially indexed, enabling fast geographic queries that weren't possible before.

UA Spoofing Detection

User-agent parsing now includes a ua_only field — the result of parsing just the UA string, ignoring any client hints that were provided.

Why does this matter? When a request sends both a UA string and client hints, comparing the two reveals spoofing. If the UA string says "Chrome on Windows" but the hints say "Safari on iOS," something is off.

resp = requests.post(
    "https://api.apicrate.io/api/v1/ua",
    headers={"X-API-Key": "YOUR_KEY"},
    json={
        "ua": "Mozilla/5.0 ... Chrome/120.0",
        "headers": {
            "Sec-CH-UA-Platform": "iOS",
            "Sec-CH-UA": "\"Safari\";v=\"17\"",
        },
    },
)

data = resp.json()["data"]
# data["client"]["name"] → "Safari" (from hints)
# data["ua_only"]["client"]["name"] → "Chrome" (from UA string alone)
# Mismatch → likely spoofed

1.8x Faster UA Parsing

We built a native Rust extension that accelerates the regex-heavy part of UA parsing. The parser attempts over 500 regex patterns sequentially to identify browsers, operating systems, and devices. The Rust RegexMatcher pre-compiles all patterns and finds the first match in a single native call, this is a great optimization that yields ~1.8x end-to-end speedup on UA parsing.

MCP Improvements

The MCP server got several upgrades:

  • Consistent naming — all tools now follow apicrate-kebab-case convention (e.g., apicrate-parse-user-agent, apicrate-check-email-risk)
  • 4 new postal tools — find-nearby-postal-codes, list-postal-systems, get-postal-system, validate-postal-codes-bulk
  • Email risk tools — check-email-risk and check-email-risk-bulk
  • Updated llms.txt — AI agents can discover all available tools automatically

Infrastructure

Under the hood:

  • PostgreSQL 18 + PostGIS 3.6 — upgraded database with spatial extensions for postal code queries
  • Connection pooling — switched from single connections to a proper async pool
  • Sphinx documentation — full API and MCP reference docs now served from the dashboard at /docs/s/

What's Next

We're working on timezone conversion, car brand/phone model datasets, and expanding the postal code database. Every new endpoint automatically becomes an MCP tool.

Sign up for free or check the full documentation to get started.

Related Posts