Window: last 3 days · dataset fibr_axp_page_hits · auto-refresh every 5 min
Bot hits
2,978
↓ 17.7% vs prior 3d
Stripped
1
→ flat · 0.0% share
Origin bot-miss
383
↑ 32.5% · 12.9% fell through
Unique paths
1,167
↓ 1.3% distinct URLs
Unique bots
14
→ flat UA matches
Avg latency
345 ms
↓ 25.1% worker wall-clock
Max latency
2,771 ms
↓ 98.3% worst-case
Daily volume stripped vs origin-bot-miss
Each bar = one UTC day. Green = served from stripped/. Orange = bot-miss (path we don't have stripped, fell through to Framer).
Daily volume by class live vs search-index vs training
Live (blue) = real-time agents fetching while a user asks an AI a question. Search-index (purple) = crawlers building a search index that AI assistants then query at chat time (Perplexity, Claude-SearchBot, etc.). Training (orange) = offline crawlers indexing for model training. Watching live + search-index trend up means humans-via-AI are finding our content.
Daily coverage % share of bot requests served from stripped/
Higher = more bot traffic served the fast, clean stripped HTML. As more pages get stripped over time, this should trend up.
Hourly heatmap day-of-week × hour, UTC
Each cell = one hour of one weekday over the 3d window. Darker = more bot hits at that time. Reveals when crawlers are most active so you can plan deploys around quiet windows. Buckets are UTC and don't follow the timezone toggle — shifting them would silently misalign the grid.
lessmorepeak: 72/hr
Latency by served variant stripped vs origin pass-through
Worker wall-clock time. Stripped = bot got pre-stripped HTML from the Assets binding. Origin-bot-miss = bot fell through to Framer (extra TLS + network round-trip). The gap is the latency benefit the strip pipeline gives bots.
Variant
Hits
Avg
Min
Max
origin-bot-miss
383
387 ms
175 ms
2,771 ms
stripped
1
429 ms
429 ms
429 ms
stripped-edited
2,594
339 ms
36 ms
1,717 ms
Status code breakdown HTTP responses we returned to bots
Healthy = mostly 2xx. Spikes in 4xx/5xx mean bots are seeing errors — investigate before they de-rank us.
Coverage by path family which sections of the site have gaps
First URL segment groups paths into families (/faq/*, /blog/*, etc.). Compare stripped vs miss to see where the strip pipeline has the worst coverage.
Family
Total
Stripped
Bot-miss
Coverage
/faq/*
751
0
0
0.0%
/blog/*
397
0
77
0.0%
/glossary/*
344
1
87
0.3%
/
296
0
0
0.0%
/landing-page/*
199
0
5
0.0%
/tools/*
179
0
3
0.0%
/conversion-rate-optimization/*
162
0
19
0.0%
/ab-testing/*
126
0
23
0.0%
/geo/*
86
0
2
0.0%
/whitepaper/*
41
0
31
0.0%
/pricing/*
33
0
0
0.0%
/about-us/*
23
0
0
0.0%
/croagency/*
18
0
1
0.0%
/pilot/*
18
0
3
0.0%
/customer-stories/*
17
0
7
0.0%
/privacy-policy/*
13
0
0
0.0%
/guides/*
12
0
2
0.0%
/seed-funding/*
12
0
0
0.0%
/partnership-program/*
10
0
10
0.0%
/compare/*
9
0
1
0.0%
By country
Code
Country
Hits
US
United States
2,253
SG
Singapore
141
PL
Poland
81
AU
Australia
71
IN
India
51
CN
China
46
DE
Germany
42
BR
Brazil
39
NZ
New Zealand
35
ES
Spain
29
JP
Japan
26
IT
Italy
23
CH
Switzerland
22
MY
Malaysia
20
KR
South Korea
19
Bot vs AS-org identity verification
Match = bot UA arrives from a network the vendor is known to use. Spoof? = UA + AS-org don't line up; likely UA fake. Unknown = no expected list yet for this bot.
Bot
AS organisation
Verdict
Hits
ChatGPT-User
Microsoft Corporation
match
709
Amazonbot
Amazon Technologies Inc.
match
602
Amazonbot
Amazon Data Services Northern Virginia
match
583
Meta-ExternalAgent
Meta Platforms Ireland Limited
match
220
ClaudeBot
Anthropic, PBC
match
181
ChatGPT-User
Microsoft Limited
match
140
Bytespider
Amazon Data Services Singapore
spoof?
140
OAI-SearchBot
Microsoft Limited
match
106
Amazonbot
Amazon.com, Inc.
match
55
Bytespider
China Unicom Hebei province network
spoof?
37
PerplexityBot
Amazon Technologies Inc.
match
36
ChatGPT-User
Microsoft do Brasil Imp. e Com. Software e Video G
match
31
Claude-User
Google LLC
match
15
GPTBot
Microsoft Limited
match
13
OAI-SearchBot
Microsoft Corporation
match
11
Perplexity-User
Amazon Technologies Inc.
match
10
Bytespider
CHINANET hebei province network
spoof?
9
GPTBot
NextGenWebs, S.L.
spoof?
5
Amazonbot
NextGenWebs, S.L.
spoof?
4
Google-CloudVertexBot
NextGenWebs, S.L.
spoof?
3
Claude-User
Broad Band Internet Service Provider, India
spoof?
3
ChatGPT-User
Microsoft Deutschland MCIO GmbH
match
3
Claude-User
IP pools
spoof?
3
ChatGPT-User
E-Light-Telecom Ltd.
spoof?
2
Claude-User
Reliance Jio Infocomm Limited
spoof?
2
Claude-User
Comcast Cable Communications, LLC
spoof?
2
Claude-User
M-net Telekommunikations GmbH
spoof?
2
ChatGPT-User
Microsoft Corp
match
2
Claude-User
AT&T Enterprises, LLC
spoof?
2
CCBot
NextGenWebs, S.L.
spoof?
2
HTTP / TLS mix spoof signal
A bot UA pretending to be a major vendor over an unusual HTTP/TLS combo (e.g. claiming to be GPTBot but speaking HTTP/1.1 + TLS 1.2 when real OpenAI uses HTTP/2 + TLS 1.3) is a likely spoof. Watch the long tail, not the dominant bar.
HTTP protocol
HTTP/21,58253.1%
HTTP/1.11,39646.9%
TLS version
TLSv1.32,93399.2%
TLSv1.2230.8%
AS-org leaderboard top networks fetching us, regardless of UA
Independent of bot name — useful for spotting volume from networks that don't claim a bot UA, or cross-checking against the Bot vs AS-org table above.
AS organisation
Hits
Distinct bots
Microsoft Corporation
721
3
Amazon Technologies Inc.
648
3
Amazon Data Services Northern Virginia
583
1
Microsoft Limited
260
4
Meta Platforms Ireland Limited
220
1
Anthropic, PBC
181
1
Amazon Data Services Singapore
140
1
Amazon.com, Inc.
55
1
China Unicom Hebei province network
37
1
Microsoft do Brasil Imp. e Com. Software e Video G
31
1
NextGenWebs, S.L.
17
7
Google LLC
15
1
CHINANET hebei province network
9
1
Microsoft Deutschland MCIO GmbH
4
2
IP pools
3
1
Live tail last 50 bot hits in the past 6h
Period-independent — same on 7D / 14D / 30D. AE write delay is ~30–90s, so "live" actually lags real-time by about a minute. Useful for incident triage and spotting bot spikes.