Reflectionbot
Few readable files block Reflectionbot in October 2026, at 22 sites or 0.4%. The community reference lists its operator as Reflection.
Who runs Reflectionbot?
| Fact | Reflectionbot reference entry |
|---|---|
| Operator | “Reflection” |
| Purpose | “Undocumented AI Agents” |
| Does Reflectionbot obey robots.txt? | “Unclear at this time.” |
| Frequency | “Unclear at this time.” |
| Cloudflare managed list for Reflectionbot | No |
These facts quote the stored community ai.robots.txt reference.
“An undocumented crawler whose user agent links to Reflection, a company that builds AI models.”
The reference leaves this agent’s purpose undocumented. Reading a page to answer a user has a different purpose from building a model with it. Owners may allow the requested visit.
How many sites block it?
Of 6,171 robots.txt files we could read, 22 block the whole site with Disallow: /, or 0.4%. Another way to read the sample is by name, with 22 files naming it in a User-agent line, or 0.4%. The 6,149 remaining files have no recorded whole-site block.
Among 552 readable files from the top 1,000 sites, we found 2 files naming it, a 0.4% share, with 2 files blocking the whole site. At 0.4%, the leading block share equals the rounded overall share. Outside this group, 5,619 files contribute 20 blocks, or 0.4%.
| Readable files | Named by | Named share | Blocked by | Blocked share |
|---|---|---|---|---|
| Reflectionbot, all ranks | 22 | 0.4% | 22 | 0.4% |
| Reflectionbot, top 1,000 | 2 | 0.4% | 2 | 0.4% |
| Reflectionbot, remaining ranks | 20 | 0.4% | 20 | 0.4% |
This is a count of blocks attached to the named crawler, covering the whole site. Blanket rules for every crawler do not enter it. Of 22 blocks, 1 does not also appear in our count of files naming it. Names and bans overlap in 21 files. We also found 1 file with a name and no whole-site ban, plus 6,149 files omitting the name. Case changes and repeated spellings are merged.
How to block or allow it
These lines go in robots.txt at the root. User-agent: * yields to a group naming Reflectionbot under the robots.txt standard.
Block the whole site
User-agent: Reflectionbot
Disallow: /Allow access
User-agent: Reflectionbot
Allow: /It's the 146th most-named crawler of the 171 we counted in the crawler index.
How we counted
For Reflectionbot, we surveyed Tranco’s top 10,000 sites on 19 October 2026. Files had an 8-second timeout and a 262,144-byte limit. Only available files we could read enter the percentages. Repeat readings are included, using scripts/robots_2026_08_06.py (robots.py at app commit 07104ff). See the full survey method.
Common questions
Does a User-agent line mean it is blocked?
This survey includes 1 file with the name and no whole-site ban, out of 22 naming files.
What do the top-ranked sites show?
Leading files record 0.4% with the name and 0.4% blocking it. Across all readable files, the corresponding shares are 0.4% and 0.4%.
Which nearby entry can I compare?
YiyanBot is named by 22 files and blocked by 22, or 0.4%. For this crawler, 22 files name it and 22 block it, a 0.4% share.