21 lookups, converters, and calculators

Domain and IP tools for records, addresses, and subnets

A domain does not have an address. It has a set of records, and each one answers a different question. Ask the wrong record and the answer looks fine while mail keeps bouncing.

This hub sorts the Toolexe network tools by the question you arrived with. Which record is wrong. Which address sits behind the name. Whether the port is listening. How many hosts fit inside a prefix. No account, no install, no terminal.

Eight record types, eight different questions

Most DNS tickets come from asking one record about another record's job. An A record knows nothing about mail. An MX record knows nothing about delegation. Keep the table below open while you work.

DNS record types, what they answer, and which Toolexe tool reads them
RecordWhat it answersWhere it goes wrongTool
AThe IPv4 address serving this nameTwo or more A records is normal. A CDN hands back a different edge per resolver.Hostname to IP
AAAAThe IPv6 address for the same nameNo AAAA means IPv6-only clients never reach the host. Nobody notices until one does.DNS Records Lookup
CNAMEThis name is an alias for another nameA CNAME at the zone apex collides with MX and NS. Providers paper over it with flattening.DNS Records Lookup
MXWhere mail for the domain is deliveredThe lower preference number wins. An MX pointing at a name with no A record bounces mail.MX Lookup
NSWhich nameservers hold the zoneDelegation lives at the registrar. Editing records on the old host changes nothing.Nameserver Lookup
SOAZone serial and refresh timersThe serial tells you whether your edit shipped. Same serial, same zone, no edit.DNS Records Lookup
TXTSPF, DKIM, DMARC, ownership proofsTwo SPF records is a hard fail, not a warning. Merge them into one string.DNS Records Lookup
PTRThe name behind an addressOwned by whoever holds the address block. Most addresses have no PTR at all.IP to Hostname

The DNS Records Lookup pulls the whole set in one pass, which beats running four separate queries and comparing them by eye. Open the single purpose tools when you already know the record and want a clean answer to paste into a ticket.

Start from the symptom, not the tool name

Nobody wakes up wanting a nameserver lookup. They wake up to a bounced invoice or a support thread. Match the row to what broke, then open the tool. The full set lives on the Domain and IP tools index.

SymptomMail bounces with "no mail server"

MX Lookup

Read the preference order first, then confirm every MX hostname resolves to a real address. A dangling MX name is the usual cause.

Check the MX chain
SymptomSite loads for you, not for a teammate

DNS Records Lookup

Compare the answer with the TTL on the record. A four hour TTL means four more hours of the old address for anyone already cached.

Read every record
SymptomMoved registrar, records look ignored

Nameserver Lookup

Delegation and zone content are two separate things. If NS still points at the old host, your new records are never asked for.

Confirm delegation
SymptomNeed the registrar, status, or expiry date

Domain Lookup

WHOIS returns registrar, status flags, and dates. Registrant contact is redacted for most domains under privacy rules.

Pull the WHOIS record
SymptomService works locally, times out from outside

Port Checker

Tests reachability from the public internet inward. Splits a firewall problem from an application problem in one attempt.

Test the port
SymptomAccess log is a wall of bare addresses

IP to Hostname

Reverse DNS names the noisy ones. Crawlers and cloud ranges show up immediately. Empty results are normal, not failures.

Resolve the address
SymptomFirewall rule needs an address, vendor gave a URL

Website to IP

Strips the scheme and path, then resolves the host. Pin rules to a hostname where the platform allows it, because CDN addresses rotate.

Resolve the URL
SymptomSupport asks which address you are coming from

What Is My IP

Shows the public address the far end sees, which is your VPN or carrier NAT exit, not the address on your laptop.

Show my address

Working an incident right now? Run the full record set first. It rules out four causes in one query and gives you something to paste into the thread.

Run a full lookup

One address, five notations

An IPv4 address is a 32 bit number wearing dots. Change the base and the same host looks like a different machine. Here is 104.21.58.189, one of the addresses answering for toolexe.com, written five ways.

Dotted quad
104.21.58.189
What people type
Decimal
1746221757
Compact storage, integer columns
Hex
68153ABD
Packet dumps, kernel structures
Octal
0150.025.072.0275
Legacy configs, filter bypass tests
Binary
01101000.00010101.00111010.10111101
Teaching masks and prefix boundaries

Two of these earn their keep beyond curiosity. Storing addresses as integers keeps range comparisons fast, so an audit query over a log table wants the decimal form. The other case is security work. A filter comparing user input against the string 127.0.0.1 waves through 0177.0.0.1 and 2130706433, and both reach loopback. Convert the input before trusting a string match.

The phone converters sit apart from the rest. They map digits into dotted notation for VoIP teaching material. No carrier is queried and no subscriber is found, so treat the output as a classroom exercise rather than a lookup result.

Size the prefix before you write the rule

Two addresses per subnet disappear into the network and broadcast slots. Forget those two and a /29 sized for six devices holds four, which surfaces on the day someone racks the seventh.

Common IPv4 prefix lengths and host counts
CIDRMaskTotalUsableTypical fit
/30255.255.255.25242Point to point link
/29255.255.255.24886A pair of appliances
/28255.255.255.2401614One rack, one service tier
/27255.255.255.2243230Small office floor
/26255.255.255.1926462Branch site
/24255.255.255.0256254The default everyone assumes
/22255.255.252.010241022Campus or VPC block

Worked example: 10.20.30.0/26

Network
10.20.30.0
First host
10.20.30.1
Last host
10.20.30.62
Broadcast
10.20.30.63
Mask
255.255.255.192
Next block
10.20.30.64/26

Sixty two usable addresses, and the next subnet starts at .64 rather than .65. Boundaries land on powers of two, never on round decimal numbers.

The CIDR Calculator prints the same breakdown for any prefix you paste, mask included. Copy the range into the firewall rule from there instead of counting octets in your head at the end of a change window.

What these tools will not tell you

A hub listing only wins is a brochure. Four honest limits, worth knowing before you quote a result to somebody else.

Geolocation places a block, not a person
Address to location data comes from registry allocations and vendor databases. It lands the right country most of the time and the right city far less often. A VPN exit or carrier NAT moves the answer hundreds of miles. Never treat it as identity evidence.
WHOIS is redacted by default
Registrar, status flags, and dates come back reliably. Registrant name and email usually resolve to a privacy service. A missing name is a policy outcome, not a hidden owner.
Reverse DNS is optional
PTR records belong to whoever holds the address block. Most public addresses have none, so an empty result from IP to Hostname means no record exists rather than a failed query.
A port check tests one path
The probe leaves our side of the internet. An open result proves a listener answered us. Your office egress rules, a proxy, or an intermediate hop still block the same port for you.

Questions from real tickets

  1. 01

    I changed an A record an hour ago and the old server still answers. What now?

    Read the TTL on the record before assuming a failure. A resolver holds the previous answer until the TTL expires, so a value of 14400 means four hours of stale replies for anyone already cached. Check the SOA serial to confirm your edit reached the zone, then wait out the timer. Drop the TTL to 300 the day before a planned move and you skip this next time.

  2. 02

    The MX record looks right and mail still bounces. Where do I look?

    An MX record names a host. The host still needs an A or AAAA record, and the receiving side still checks your TXT records. Two SPF strings on one domain is a hard fail. A DKIM selector dropped during a migration fails quietly for weeks. Pull the full record set, not the MX row alone.

  3. 03

    Is 0177.0.0.1 a real address?

    It resolves to 127.0.0.1. Octal notation with a leading zero, decimal notation like 2130706433, and hex notation all reach loopback, and plenty of allowlist filters compare strings instead of parsing the address. Convert the value first, then decide whether the input belongs anywhere near an outbound request.

  4. 04

    Two lookups of one domain returned different addresses. Is something broken?

    Usually no. A domain holds several A records, and anycast networks answer from the edge nearest the resolver asking. A lookup run from our servers reports the edge we reached. Treat a single answer as one sample, not the map.

  5. 05

    Does an open port mean the service is reachable from my office?

    No. A port check reaches the host from one place on the internet. Your office egress rules, a corporate proxy, or a hop in between still block the same port for you. Open means the listener exists and answered us. It stops short of proving your path works.

  6. 06

    Do these lookups leave a trace?

    Public DNS and WHOIS are public systems, and the resolvers on the far end keep their own logs. Nothing you look up here stays private from them. Treat a lookup box the way you treat a query typed into a shared terminal.

Where to go after the record checks out

DNS answers stop at the edge of the host. Certificate chains, headers, and hashes sit in the security set. Encoding a payload for transport sits elsewhere again.

Read the record before blaming the server. Read the TTL before blaming the record. Convert the address before trusting a string comparison.