IP Ranges

Continuously-updated public IP ranges for major cloud providers, services, and apps — one txt file per provider, ready to drop into firewalls, allowlists, or geo blocks.

35 providers · 113,771 IPv4 entries · 267,505 IPv6 entries · refreshed every 4 hours via GitHub Actions

All addresses come from public sources (vendor-published JSON/TXT or DNS resolution of vendor domains). The lists are committed back to main after each refresh, so consumers can pin to either main (rolling) or a specific commit (frozen).

Quick start

Each provider directory contains four files:

File	Contents
ipv4.txt	IPv4 CIDRs, one per line
ipv4_merged.txt	Same list reduced to the smallest equivalent set of CIDRs
ipv6.txt	IPv6 CIDRs, one per line
ipv6_merged.txt	Same list reduced

Fetch them straight from raw.githubusercontent.com:

curl -fsSL https://raw.githubusercontent.com/mrkhachaturov/ipranges/main/cloudflare/ipv4_merged.txt

Use in ipset + iptables

ipset create cloudflare hash:net family inet
curl -fsSL https://raw.githubusercontent.com/mrkhachaturov/ipranges/main/cloudflare/ipv4_merged.txt \
  | xargs -I{} ipset add cloudflare {}
iptables -I INPUT -m set --match-set cloudflare src -j ACCEPT

Use in nginx

geo $is_googlebot {
    default 0;
    include /etc/nginx/googlebot-ipv4.conf;  # generated from googlebot/ipv4_merged.txt as "<cidr> 1;"
}

Use in a Python allowlist

import ipaddress, urllib.request
url = "https://raw.githubusercontent.com/mrkhachaturov/ipranges/main/openai/ipv4_merged.txt"
nets = [ipaddress.ip_network(line.strip()) for line in urllib.request.urlopen(url) if line.strip()]
def is_openai(ip: str) -> bool:
    addr = ipaddress.ip_address(ip)
    return any(addr in n for n in nets)

Providers

Provider	IPv4	IPv4 (merged)	IPv6	IPv6 (merged)	Notes
Akamai	txt	txt	txt	txt
Amazon (AWS)	txt	txt	txt	txt
Anthropic (Claude)	txt	txt	txt	txt
Apple (Private Relay)	txt	txt	txt	txt
Atlassian	txt	txt	txt	txt
Bing (Bingbot)	txt	txt	—	—
Cloudflare	txt	txt	txt	txt
Cloudflare Tunnel (Argo)	txt	txt	txt	txt
Amazon CloudFront	txt	txt	txt	txt
DeepL	txt	txt	txt	txt
Devolutions	txt	txt	—	—
DigitalOcean	txt	txt	txt	txt
Discord	txt	txt	—	—
Facebook (Meta)	txt	txt	txt	txt
Games	txt	txt	txt	txt
GitHub	txt	txt	txt	txt
Google (Cloud & GoogleBot)	txt	txt	txt	txt
Google (GoogleBot)	txt	txt	txt	txt	To allow GoogleBot, block all Google IPs first, then allow these.
Groq	txt	txt	txt	txt
Kino.pub	txt	txt	txt	txt
Linode	txt	txt	txt	txt
Microsoft	txt	txt	txt	txt
Notion	txt	txt	txt	txt
OpenAI (GPTBot)	txt	txt	txt	txt
OpenTofu	txt	txt	txt	txt
Oracle (Cloud)	txt	txt	—	—
ProtonVPN (exit nodes)	txt	txt	—	—
Roblox	txt	txt	txt	txt
Spotify	txt	txt	txt	txt
Sunsama	txt	txt	txt	txt
Tana	txt	txt	txt	txt
Telegram	txt	txt	txt	txt
Twitter / X	txt	txt	txt	txt
Vultr	txt	txt	txt	txt
Wispr Flow	txt	txt	txt	txt
All-in-one (every provider combined)	txt	txt	txt	txt	Aggregate of every provider above

The provider table above is generated from utils/providers.json by utils/render_readme.py. Don't edit between the <!-- BEGIN AUTO:* --> markers by hand — the GitHub Action will overwrite changes on the next run.

How updates work

A GitHub Action runs every 4 hours (cron: '8 */4 * * *') and on workflow_dispatch:

1. Executes every */downloader.sh discovered under the repo
2. Concatenates all per-provider lists into all/ipv4.txt and all/ipv6.txt
3. Runs utils/merge.py to produce *_merged.txt for each list
4. Runs utils/render_readme.py to refresh the table and counts above
5. Commits any changes back to main

If you want fresher data than the cron, trigger the Update workflow manually from the Actions tab.

Adding a new provider

The full contract lives in CLAUDE.md. The short version:

1. Create <provider>/downloader.sh. Use devolutions/downloader.sh (DNS-resolution pattern) or google/downloader.sh (vendor-JSON pattern) as templates. Both invocation modes — repo root and inside the provider directory — must work.
2. Run it once locally to seed <provider>/ipv4.txt and <provider>/ipv6.txt.
3. Add an entry to utils/providers.json (kept in alphabetical order). The README updates itself on the next workflow run.

Local development

# system deps
sudo apt install -y whois parallel gawk dnsutils jq python3-pip

# python deps
pip install -r utils/requirements.txt

# run a single provider
cd cloudflare && bash downloader.sh

# regenerate the merged file for one list
python utils/merge.py --source=cloudflare/ipv4.txt | sort -V > cloudflare/ipv4_merged.txt

# refresh README from current data
python utils/render_readme.py

# CI guard — fail if README would change
python utils/render_readme.py --check

License

MIT — © 2026 Ruben Khachaturov. The IP data itself comes from public sources and is not subject to copyright on its own.

Source

https://github.com/mrkhachaturov/ipranges