IP & Subnet Tools
Enter inclusive endpoints or a list of ranges. Get the smallest exact CIDR set without adding addresses.
Addresses are processed in this browser; no network lookup or upload is performed.
Exact coverage only. Overlapping and adjacent ranges are merged within each IP version; gaps stay excluded.
One IP or start-end per line. Use an ASCII hyphen; blank lines are ignored.
Up to 100 nonempty ranges and 65,536 input characters. Results appear 100 blocks at a time.
Input changed. Convert again before exporting.
IP range to CIDR helps when an allowlist gives you first and last addresses but the destination accepts prefixes. A convenient single supernet can quietly include addresses outside that list.
Start with the nonaligned example to see where prefix boundaries fall. For batches, compare the input range count with the merged interval count before taking the result into a rule editor.
Keep the intended address set visible while changing its notation.
IP range to CIDR splits a nonaligned interval into several blocks instead of expanding it. Each block shows its first address, last address and exact size so you can check the edges.
Two overlapping ranges should not produce duplicate rules. The batch engine unions ranges of the same version, keeps disjoint gaps and then calculates the minimal CIDR blocks for that union.
The list accepts one address or start-end per nonempty line. Both endpoints are inclusive; use an ordinary hyphen rather than a typographic dash.
Auto conversion waits 350 ms after an edit. Turning it off cancels pending work and reveals the convert button.
The screen adds 100 blocks at a time. Copy and both downloads always contain every block, including those not yet shown.
The IP range converter accepts strict dotted IPv4, compressed IPv6 and IPv4 embedded in IPv6. IPv6 range to CIDR uses integer arithmetic throughout, including values larger than JavaScript’s safe integer range.
Duplicate, overlapping and adjacent input intervals are merged separately for IPv4 and IPv6. Output is sorted by address, with IPv4 first; no output block straddles a gap.
Counts include every address, without subtracting network or broadcast addresses. A /31 has two addresses, a /32 or /128 has one, and the IPv6 /0 count is kept intact in JSON.
CIDR blocks must start on a power-of-two boundary and have a power-of-two size. An interval such as 192.0.2.5 through 192.0.2.30 does not satisfy those conditions as a single block, so it needs several exact blocks.
For the same address set, the generated block count is minimal. Further reduction would require adding addresses or dropping part of the input. Batch conversion first merges ranges so repeated coverage does not inflate the result.
Yes. Full and compressed IPv6 endpoints are accepted, including very large intervals. Output stays IPv6, even for IPv4-mapped input, and address counts use exact decimal strings.
The conversion stops with an error. Endpoints are never automatically swapped: correct the range explicitly, then run it again. In a batch, no valid-looking partial output can be exported.
A single inclusive range needs at most 62 IPv4 blocks or 254 IPv6 blocks. Those bounds occur just inside the full address-space edges. The full space itself needs only one /0 block.
Still have a question?
Use IP range to CIDR with your exact endpoints, check the boundary addresses, and download the complete set in the format your next step needs.