PerplexityBot IP Ranges
Perplexity runs two robots: PerplexityBot, which indexes pages for Perplexity's answer engine, and Perplexity-User, which fetches a page when someone asks Perplexity about it. Each has its own address list. Together they come to 12 IPv4 prefixes and only 32 addresses, the smallest footprint of the crawlers on this site. Perplexity owns no network for them: every prefix resolves to Amazon's AS14618, so WHOIS on these addresses names Amazon.
- Provider
- PerplexityBot
- Hosting network
- AS14618 (Amazon AWS, host)
- Category
- Crawler
- Headquarters
- San Francisco, CA, USA
- Published IPv4 prefixes
- 12
- Registry
- Self-published list
Known IP ranges
Perplexity does not announce these addresses under an ASN of its own. They are the prefixes it publishes in https://www.perplexity.com/perplexitybot.json, https://www.perplexity.com/perplexity-user.json, as fetched on 2026-10-16; the space itself is routed by AS14618 (Amazon AWS). The list decides whether an address belongs to PerplexityBot - WHOIS for any of them returns the host network.
18.97.9.96/29
18.97.43.80/29
18.97.1.228/30
18.97.21.0/30
3.211.124.183/32
3.222.232.239/32
3.224.62.45/32
3.231.139.107/32
18.210.92.235/32
34.193.163.52/32
What does a PerplexityBot IP mean in a privacy test?
A Perplexity address is a robot fetching a page for an AI answer, not a visitor. With only 32 addresses in total, a request outside the two lists that claims to be PerplexityBot is almost certainly something else.
Two lists, mostly single addresses
The files are www.perplexity.com/perplexitybot.json and www.perplexity.com/perplexity-user.json. Their creationTime stamps read 7 February 2025 and 17 October 2025 when fetched on 16 October 2026.
Where Google hands out /27s, Perplexity lists mostly individual hosts. PerplexityBot's eight entries are six single addresses, one /29 and one /30, starting with 3.211.124.183/32; Perplexity-User's four are two single addresses, a /29 and a /30. Merged, the two lists cover 32 addresses, in AWS space under 3.0.0.0/8, 18.0.0.0/8 and a few other blocks.
Indexing versus answering a user
The two robots do different jobs, so the list an address appears in matters. PerplexityBot builds the index Perplexity answers from, and is the one robots.txt rules are written for. Perplexity-User fetches a page because a person asked about it while using Perplexity, so its visits follow what users ask rather than a crawl schedule.
Because both run on AWS, the network name proves nothing on its own: AS14618 also carries Common Crawl's CCBot and a very large number of ordinary AWS customers. Matching the address against the two lists is the test.
Related tools
Frequently asked questions
What IP ranges does PerplexityBot use?
The eight entries in perplexity.com/perplexitybot.json: six single addresses, a /29 and a /30, all on Amazon's AS14618. Perplexity-User has four more entries in a list of its own.
Why does a Perplexity address appear in my logs or a privacy test?
A Perplexity address is a robot fetching a page for an AI answer, not a visitor. With only 32 addresses in total, a request outside the two lists that claims to be PerplexityBot is almost certainly something else.
Why does WHOIS say Amazon for a Perplexity address?
Perplexity runs its crawlers on AWS and publishes the addresses it uses; the registry names the network owner. Every prefix in both lists resolved to Amazon's AS14618 on 16 October 2026.
What is the difference between PerplexityBot and Perplexity-User?
PerplexityBot indexes pages for Perplexity's answers; Perplexity-User fetches a page when a person asks Perplexity about it. Each has its own user-agent and its own address list.
How many addresses does Perplexity crawl from?
Thirty-two, across the two lists combined as of October 2026. That is small enough to allow or block address by address.