Thanks - the HTTP point is correct, and I'm fixing it: ipinfo moves to HTTPS, and ip-api's free endpoint can't do TLS, so it goes or gets replaced with a keyed source. On host type, I agree it's more decision-relevant than a fifth city. I'm adding a structured classification built from what the tool already has — AbuseIPDB usageType/isTor, RDAP/WHOIS, and ASN — with the evidence and confidence attached to each flag, and an explicit "unknown" for VPN/proxy where no free source exists. I am not currently using it for work at this time. I'll keep IPGeolocation.io in mind as an optional keyed source, though I'd keep paid host-type flags out of the default path. Appreciate the disclosure.
