Both IPv4 addresses (validated, dotted-quad) and IPv6 addresses including compressed forms. Each is replaced with a bracketed placeholder so the log structure stays intact.

What do consistent tokens do?

With consistent tokens on, the same IP always maps to the same placeholder like [IPv4_3], so you can still group requests by client during analysis. With it off, every address becomes a generic [IPv4] or [IPv6].

Does my log data leave the browser?

No. All matching runs in your browser. Nothing is uploaded, logged, or stored.

Are IP addresses really personal data?

Yes. Under the GDPR, IP addresses are treated as personal data because they can identify an individual, which is exactly why you should strip them before sending logs to a third-party AI service.

Can it misfire on version numbers or other numbers?

IPv4 detection is strictly validated to 0–255 octets, so most version strings are safe. IPv6 matching is broad by design, so glance over the output if your data contains many colon-separated hex values.

What is the IP Address Scrubber?

Paste server logs or network data and automatically detect and replace every IPv4 and IPv6 address with anonymized placeholders — preserving log structure for AI analysis without exposing user IP addresses, entirely in your browser. It runs free in your browser on Gera Tools, with nothing uploaded.

IP Address Scrubber — Gera Tools

Name: IP Address Scrubber
Creator: Gera Tools
License: https://creativecommons.org/licenses/by/4.0/

Get one useful tool a week

Like this tool? Enter your email and we'll send you one genuinely useful Gera tool a week — plus a link to come back to this one. No spam, one-click unsubscribe any time.

Why scrub IP addresses before AI analysis?

Server logs are gold for AI-assisted debugging — and full of personal data. Under the GDPR and most modern privacy frameworks, IP addresses are classified as personal data because they can identify a specific individual or device. Pasting raw logs into a third-party AI model means those IP addresses leave your environment and may be stored or processed by the provider. The safe path is to anonymize first, analyze second.

This tool detects every IPv4 and IPv6 address in your pasted log text and replaces each one with a bracketed placeholder, entirely in your browser, before you paste the result into any AI tool.

How it works

Paste your logs and the tool runs two detection passes:

IPv4 detection uses a strictly validated pattern: each octet must be a number between 0 and 255, and the address must have exactly four parts. This prevents common false positives like software version strings (3.11.2) or timestamps (2024.01.15) from being caught.
IPv6 detection uses a broader pattern that covers compressed forms with :: and full eight-group addresses.

With consistent tokens enabled, each unique IP address maps to a stable numbered placeholder — [IPv4_1], [IPv4_2], [IPv6_1] and so on. This means the same source address always gets the same token throughout the log, so you can still correlate multiple requests from the same client during analysis. With consistent tokens off, every address becomes a flat [IPv4] or [IPv6] with no identity preserved.

A count of unique IPv4 and IPv6 addresses replaced confirms what was found in the text.

What gets replaced, and what does not

The scrubber targets IP address patterns specifically. Log lines like this:

192.168.1.45 - - [15/Jan/2024:10:23:01 +0000] "GET /api/users HTTP/1.1" 200 4521

become:

[IPv4_1] - - [15/Jan/2024:10:23:01 +0000] "GET /api/users HTTP/1.1" 200 4521

The timestamp, request method, status code, response size, and path are untouched. Only the IP address is replaced, keeping the log structure fully intact for analysis.

Consistent tokens vs. flat redaction: when to use each

Mode	Output	Use when
Consistent tokens ON	`[IPv4_3]` stable per address	You need to track same-client requests
Consistent tokens OFF	`[IPv4]` generic for every match	You only care about log structure, not identity

For most debugging scenarios — finding a 500 error pattern, analyzing slow requests, understanding redirect chains — consistent tokens are more useful because they preserve the ability to say “these three requests came from the same user.”

Tips and notes

IPv4 is tightly validated. Octets must be 0–255, so version strings and most ordinary numbers are left alone.
Glance at IPv6 output. That pattern is intentionally broad — if your data has lots of colon-separated hex values for other purposes, confirm nothing benign was caught.
Nothing is uploaded. All pattern matching runs in your browser. Your log data never leaves your device.