A dependency-free IP allowlist/blocklist. Think Node.js net.BlockList, but faster.
Accepts IPv4 and IPv6 addresses, ranges, and subnets, and uses binary search for fast lookups.
npm install ip-range-listIn client-side applications, you can import the package directly via jsDelivr CDN.
Using the package in ESM/CJS applications:
import { IPRangeList } from 'ip-range-list';
let ranges = new IPRangeList()
.addRange('192.0.1.1', '192.0.1.255')
.addSubnet('192.0.2.9/24')
.addAddress('2001:db8::1');
ranges.contains('192.0.2.42'); // => true
ranges.contains('::ffff:192.0.2.42'); // => true
ranges.contains('2001:db8::2'); // => falseUsing the UMD build:
const { IPRangeList } = window.ipRangeList;
let ranges = new IPRangeList();Note
IPv4-mapped IPv6 addresses such as ::ffff:192.0.2.1 are treated as equivalent to plain IPv4 addresses such as 192.0.2.1 across all methods.
Creates an empty mutable list.
Adds one IPv4 or IPv6 address and returns the same list.
Adds the inclusive range between two addresses and returns the same list.
Adds a CIDR subnet and returns the same list. Host bits are normalized, so 192.0.2.9/24 adds 192.0.2.0 through 192.0.2.255.
Return whether an address belongs to a stored range. .check() is an alias for .contains(). Invalid candidates return false.
A read-only snapshot of the currently stored canonical merged ranges:
[
{
start: '::ffff:192.0.2.0',
end: '::ffff:192.0.2.255',
},
]- Overlapping and adjacent ranges merge into one canonical inclusive interval.
- IPv4 is stored as IPv4-mapped IPv6 (
::ffff:0:0/96), so dotted IPv4 and mapped IPv6 candidates compare identically. - Public mutation methods reject malformed addresses and unsupported input types with
TypeError. - Out-of-range CIDR prefix lengths and reversed ranges throw
RangeError.
Lookups use binary search over the canonical ranges.
The package includes a benchmark comparing IPRangeList with Node.js net.BlockList against
Manycast IPv4 and IPv6 prefix data. See the benchmark notes for the
data download script, query profiles, methodology, and reproduction steps. The published run used 55,473 prefixes:
41,207 IPv4 prefixes and 14,266 IPv6 prefixes.
The present profile checks addresses from the dataset prefixes, the missing profile checks generated addresses
outside the loaded ranges, and the mixed profile uses 20% present addresses and 80% missing addresses.
| Scenario | Profile | ip-range-list |
node:net BlockList |
Result |
|---|---|---|---|---|
| Full import | - | 54.09 ms | 40.36 ms | BlockList 1.34x faster |
| 1,000,000 lookups after import | present | 1,472.24 ms | 184,395.46 ms | ip-range-list 125.25x faster |
| 1,000,000 lookups after import | missing | 1,568.54 ms | 394,476.87 ms | ip-range-list 251.49x faster |
| 1,000,000 lookups after import | mixed | 1,609.40 ms | 356,680.64 ms | ip-range-list 221.62x faster |
| Interleaved import/lookups | present | 71.01 ms | 1,009.32 ms | ip-range-list 14.21x faster |
| Interleaved import/lookups | missing | 96.28 ms | 2,023.87 ms | ip-range-list 21.02x faster |
| Interleaved import/lookups | mixed | 98.54 ms | 1,818.27 ms | ip-range-list 18.45x faster |
Refer to the contributing guide.