Monid · monid.ai · Last 24 hours
7 hoppers · 8 sends
A hopper is a daily HMAC nick, not a person. Read Path A vs B/C next — only browser-like hops can later join a buy in PostHog.
How to read this page
- Hop — a 302 we issued. This is the only send we can prove.
- Impression — they saw our listing. Not a hop, not a buyer.
- Daily nick — HMAC(day, IP, UA). Rotates at 00:00 UTC. Not a person.
- Path A / B / C — A is a human-looking click. B is a tool fetch then a later human. C is the agent as buyer. We infer A vs B/C from headers. We do not store IP or raw User-Agent.
302s we issued. The only send we can prove.
HMAC(day, IP, UA). Rotates at 00:00 UTC. Not a person.
Same nick hopping more than once in a UTC day.
0 nicks saw an index/MCP listing. A saw, not a hop.
Do not read impressions → hops as a funnel. Most hops come from listings we do not see as a saw, so hops can be larger.
Who they look like when they hop
37.5% of hops could be either a browser or a tool. 0.0% look like a human/browser click (Path A). 62.5% look like an agent/tool fetch (Path B or C).
Looks like a human / browser click
0Sec-Fetch-Dest=document or a browser UA. If a person opened the hop, PostHog can later join a buy.
Looks like an agent / tool fetch
5curl, python, JSON Accept, or a CORS fetch. The tool read the page. A later human buy is usually unjoinable unless they keep utm_id.
Headers were not enough to tell
3Still a real hop. We do not guess. Role is a proxy, not a person label.
How they found you
They find the hop most through Publisher network (100.0%). They open Remote MCP most often.
Channel is the listing that held the hop (utm_term). Dest is the page or file the 302 pointed at (utm_content).
Channel
Where the agent discovered the hop.
-
1
Publisher network8 100.0% · 7 nicks
Destination on your site
Which of your landings we sent them to.
-
1
Remote MCP5 62.5% · 4 nicks
-
2
Skill file2 25.0% · 2 nicks
-
3
MCP setup guide1 12.5% · 1 nicks
Can their PostHog see it?
PostHog can see this hop if their JS is on the page.
Markdown, text, or install. PostHog JS will not fire. We still count the hop.
Machine endpoint. No pageview. A buy only joins if they copy utm_id onto the key or invoice.
Immediate login/signup redirect. Query is dropped unless they persist utm_id first.
0.0% of hops landed on HTML. Only those can appear in the client PostHog project unless they server-capture the query string.
Where the request came from
Cloudflare saw United States (US) most often (75.0%). That is the hopper, not a buyer address.
Country
ISO country of the hop request.
-
1
United States (US)6 75.0% · 5 nicks
-
2
Finland (FI)2 25.0% · 2 nicks
Cloudflare colo
Edge city that saw the 302.
-
1
ORD5 62.5% · 4 nicks
-
2
WAW2 25.0% · 2 nicks
-
3
IAD1 12.5% · 1 nicks
Continent
-
1
NA6 75.0% · 5 nicks
-
2
EU2 25.0% · 2 nicks
Network family
-
1
google5 62.5% · 4 nicks
-
2
hetzner2 25.0% · 2 nicks
-
3
microsoft1 12.5% · 1 nicks
What the operator declared
bing leads at 1 hops (12.5%).
Allowlisted families only. Raw User-Agent is discarded after we pick a family.
UA family
-
1
none7 87.5% · 6 nicks
-
2
bing1 12.5% · 1 nicks
Accept-Language
Primary language class (en, zh, …). Not a person name.
-
1
und8 100.0% · 7 nicks
Agent class
verified_agent / declared_ai / likely_agent.
-
1
likely_agent8 100.0% · 7 nicks
How we decided
declared UA, machine endpoint, signed, or unknown.
-
1
machine_endpoint8 100.0% · 7 nicks
Bot score band
Bot score was not present on these hops.
Kind
browser vs api, from Accept and path.
-
1
api5 62.5% · 4 nicks
-
2
browser3 37.5% · 3 nicks
How they asked for the hop
json leads at 5 hops (62.5%).
These classified headers are the best proxy for browser vs tool. We never store the raw strings.
Accept
html = wants a page. text/json = usually a tool.
-
1
json5 62.5% · 4 nicks
-
2
wildcard3 37.5% · 3 nicks
Fetch dest
document = browser navigation. empty/cors = usually a fetch.
-
1
empty8 100.0% · 7 nicks
HTTP version
-
1
h26 75.0% · 5 nicks
-
2
h12 25.0% · 2 nicks
TLS
-
1
tls137 87.5% · 6 nicks
-
2
unknown1 12.5% · 1 nicks
When the hops happened
UTC send time of the 302. This is not purchase time. 7d drops days older than a week.
Hops by day
Peak 10-23: 4 hops
Calendar day of the 302.
Hour of day (UTC)
Busiest hour 02:00 UTC — 3 hops
Send time of the 302, not purchase time. Darker is more hops.
- 00
- 01
- 02 3
- 03
- 04
- 05 1
- 06 2
- 07
- 08 1
- 09
- 10
- 11 1
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
Recent hops
Latest classified 302s. Nick is eight characters of the daily hash. Filter by how the request looked.
| When (UTC) | Looks like | Opened | Found via | Where | Operator | Asked for | Nick |
|---|---|---|---|---|---|---|---|
| 2026-10-24 05:30:38 | Likely agent / tool fetch | Remote MCP api | publisher | United States (US) · ORD | none | json / empty | 21260625 |
| 2026-10-24 02:39:36 | Could be either | Skill file agent_fetch | publisher | Finland (FI) · WAW | none | wildcard / empty | 20ae0442 |
| 2026-10-24 02:31:28 | Could be either | MCP setup guide agent_fetch | publisher | United States (US) · IAD | bing | wildcard / empty | d12dca0a |
| 2026-10-24 02:00:29 | Likely agent / tool fetch | Remote MCP api | publisher | United States (US) · ORD | none | json / empty | fe12bad4 |
| 2026-10-23 11:06:13 | Could be either | Skill file agent_fetch | publisher | Finland (FI) · WAW | none | wildcard / empty | e35d60dc |
| 2026-10-23 08:34:51 | Likely agent / tool fetch | Remote MCP api | publisher | United States (US) · ORD | none | json / empty | a4e3584b |
| 2026-10-23 06:53:56 | Likely agent / tool fetch | Remote MCP api | publisher | United States (US) · ORD | none | json / empty | 2405c4fd |
| 2026-10-23 06:52:26 | Likely agent / tool fetch | Remote MCP api | publisher | United States (US) · ORD | none | json / empty | a4e3584b |