Reflectionbot (Reflection AI): what it is and how to allow or block it
Updated · checked against the operator's documentation on
Reflectionbot is the web crawler of Reflection AI. Reflection AI says it visits publicly available pages to help build and improve its models, reads robots.txt and honors it. Block it with User-agent: Reflectionbot. Its robots.txt token is Reflectionbot.
Seen Reflectionbot/1.0 in your logs?
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Reflectionbot/1.0; +https://reflection.ai/bot) Chrome/151.0.0.0 Safari/537.36The example string from Reflection AI's documentation. Version numbers change; the token Reflectionbot stays.
- What it is
- An AI training crawler run by Reflection AI.
- Honors robots.txt
- YesYes, per Reflection AI.
- Real or fake?
- Can be verifiedThe user agent is only a claim. Check the requesting IP against Reflection AI's published IP ranges.
- If you block it
- Reflection AI stops visiting your pages for its models.
Facts
- Operator
- Reflection AI
- Category
- AI training crawler
- robots.txt token
Reflectionbot
- Honors robots.txt?
- Yes, per Reflection AI. Reflection AI says Reflectionbot reads robots.txt at the site root and honors it, and recommends robots.txt over blocking with HTTP 403. It says changes are typically picked up within about a day. Match on the token Reflectionbot, not the full string.
- User-agent example
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Reflectionbot/1.0; +https://reflection.ai/bot) Chrome/151.0.0.0 Safari/537.36
- Official source
- Reflection AI: Reflectionbot
- Published IP ranges
- https://reflection.ai/bot/ips.json
- Compared with the source
What it is used for
Per Reflection AI, Reflectionbot visits publicly available pages to help build and improve the company's models. It says this is the only user agent it crawls with today.
Collects public pages that may be used to train AI models. Blocking it is a statement about training, not about being found.
What happens if you block it
Reflection AI stops visiting your pages for its models. Nothing about your search visibility changes; it is not a search engine crawler.
Reflection AI says genuine requests come from IP addresses that resolve by reverse DNS to hostnames under reflection.ai and publishes its ranges at https://reflection.ai/bot/ips.json. The ranges can change, so compare with the current file.
Is it really Reflectionbot?
Anyone can send Reflectionbot as a user agent, so a log line proves nothing by itself. The way to tell is the requesting IP address.
- Compare the IP address with the Reflection AI: published IP ranges. A request from outside those ranges is not the operator's crawler.
- Guide: robots.txt syntax: groups, wildcards and matching rules.
- Guide: Fake Googlebots and PHP scanners: what hit our new server in 48 hours.
- Paste any user agent from your logs into the user-agent lookup, or follow the scripts in Fake Googlebots and PHP scanners.
How to allow or block Reflectionbot
Add one of these groups to the robots.txt at the root of your domain. A group that names the bot overrides a general User-agent: * group.
User-agent: Reflectionbot Allow: /
User-agent: Reflectionbot Disallow: /
Block Reflectionbot at the server
robots.txt asks; a server rule refuses. These answer 403 to any request whose user agent contains reflectionbot (case-insensitive). They stop only bots that name themselves honestly, so keep the robots.txt group as well.
# server { ... }
if ($http_user_agent ~* "Reflectionbot") {
return 403;
}RewriteEngine On
RewriteCond %{HTTP_USER_AGENT} Reflectionbot [NC]
RewriteRule .* - [F,L]Step by step, with a test command: Block AI bots in nginx by user agent · Block AI bots in Apache with .htaccess · Block AI crawlers on Cloudflare · Block AI bots on Vercel
Does your site allow Reflectionbot today? The checker evaluates your robots.txt for this and every other bot.Check my site
What we see on this site
Live counter, last 30 days: 33 requests on 11 days identified as Reflectionbot (latest: ). That is a small sample: read it as an example, not as a measurement of the web.
- Content pages: 25
- robots.txt: 7
- Images and files: 1
Counts come from this site's own server; they hold no IP addresses or URLs, and a user agent is only a claim. See the bot traffic data and the methodology.
Questions about Reflectionbot
What is Reflectionbot?
Reflection AI's crawler that visits public pages to help build and improve its models. Per Reflection AI, Reflectionbot visits publicly available pages to help build and improve the company's models. It says this is the only user agent it crawls with today.
Is Reflectionbot safe and legitimate?
The real Reflectionbot is an AI training crawler run by Reflection AI. A user-agent string is only a claim, though, so anyone can send it. Compare the requesting IP address with the published IP ranges.
Does Reflectionbot obey robots.txt?
Yes, per Reflection AI. Reflection AI says Reflectionbot reads robots.txt at the site root and honors it, and recommends robots.txt over blocking with HTTP 403. It says changes are typically picked up within about a day. Match on the token Reflectionbot, not the full string.
How do I block Reflectionbot?
Add a group to the robots.txt at the root of your domain: User-agent: Reflectionbot followed by Disallow: /. A group that names the bot overrides a general User-agent: * group.
Will blocking Reflectionbot hurt my Google rankings or AI visibility?
Blocking it is a statement about training, not about being found in search. Reflection AI stops visiting your pages for its models. Nothing about your search visibility changes; it is not a search engine crawler.
What does "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Reflectionbot/1.0; +https://reflection.ai/bot) Chrome/151.0.0.0 Safari/537.36" mean in my logs?
It is the user-agent string in Reflection AI's documentation for Reflectionbot, an AI training crawler run by Reflection AI. The part that robots.txt rules match is the token Reflectionbot. The string is only a claim, because anyone can send it; the section "Is it really Reflectionbot?" above shows how to check a request.
Related bots
Other AI training crawlers
- GPTBot OpenAIOpenAI's crawler for public web content that may be used to train its generative AI models.
- ClaudeBot AnthropicAnthropic's crawler that collects public web content which may contribute to training its generative AI models.
- CCBot Common CrawlThe crawler of Common Crawl, a non-profit that maintains an open repository of web crawl data.
- meta-externalagent MetaMeta's crawler that gathers web content to train AI models or to improve products through indexing.
All AI Training Crawlers · Bot directory · Which bots to allow or block · robots.txt generator · See a page as a bot