All numbers below are production traffic unless noted. dev rows
(preview/branch & CI pods) are excluded from user-behavior sections. Two surfaces exist: the
map-first Explore / Chiku page (where nearly all traffic is) and a separate
pservs-backed /search page.
Traffic is dominated by the Explore / Chiku page. The funnel is visible in the event mix: people set a location, Chiku suggests destinations, places get resolved, and some convert to a trip search.
| Event | Count | What it is |
|---|---|---|
llm.prompted | 334 | A user message to Chiku |
places.resolved | 233 | Origin/destination resolved via WAPI |
places.predictions | 193 | Address autocomplete calls |
trips.searched | 169 | Actual trip search (MCP) |
places.details | 58 | Place detail fetches |
origin.detected | 34 | IP-based "use my location" |
search.completed | 8 | /search page (dev only) |
errors | 4 | 3× WAPI auth + 1× LLM |
Haiku is comfortably handling the task (pick a few destinations + write two sentences). Cost is
negligible, latency is reasonable, and the model effectively never returned an off-list destination.
Active system prompt: chiku-system-base-v5.
315 production messages, grouped by primary intent (a message can touch more than one theme).
| Intent | ~Msgs | Examples |
|---|---|---|
| Setting a location | ~140 | 84 start with “From …”; bare names (omaha, sf); routes (denver to cheyenne) |
| Vibe / inspiration | ~26 | somewhere warm, Beach weekend, Mountain town, one day trip, foodies |
| Date / timing | ~24 | In August, august 24-27, cheap day trip … today |
| Vague follow-ups | ~24 | More, Anywhere, Yes, What else, Idk |
| Budget-led | ~19 | Cheapest, Round trip under 50, $100 and a free july… |
| Nearest stop / station | ~16 | where will it stop in dc, closest stop to long island |
| Train / Amtrak mode | ~14 | Where can I go by Amtrak from St. Paul, does a massive train go to PA |
| Off-network / international | ~10 | Berlin, Copenhagen, France, Nayarit, northern canada |
| Logistics / policy | ~9 | Traveling with service dog, Does a 6 year old need a ticket, senior fares |
| Support / UI frustration | ~7 | see callouts below |
Location-setting plus open-ended vibe/budget inspiration dominate — exactly what the explore surface is for. “Where am I → somewhere warm / cheap” is the dominant pattern.
Nearest-stop / station detail (~16 msgs) — users want to know boarding points. Amtrak/rail-specific routing (~14) — one St. Paul user rephrased the same rail question 4 times, suggesting Chiku wasn’t answering it.
“Hello I’m waiting on my bus … haven’t seen no Wanda bus where is my bus”
“Just give me the options like old pen I don’t wanna see the ai”
Lost-passenger, refund, and policy questions land on Chiku with no handoff. A vocal minority
wants a plain map/list instead of a chatbot — relevant for the inspiration-page experiment readout.
From the 169 trip searches. Destinations come from a fixed seed list (only 16 ever appear); origins are free-text / geolocated, so they sprawl.
| Destination | Searches |
|---|---|
| New York, NY | 24 |
| Chicago, IL | 24 |
| Atlanta, GA | 15 |
| Miami, FL | 14 |
| Providence, RI | 12 |
| Philadelphia, PA | 11 |
| Washington, DC | 10 |
| Los Angeles, CA | 10 |
| Montreal, QC | 9 |
| Boston, MA | 9 |
| San Francisco, CA | 8 |
| Las Vegas, NV | 8 |
| Cleveland · Detroit · Pittsburgh · Baltimore | 5 / 4 / 3 / 3 |
| Origin | Searches |
|---|---|
| New York, NY (+variants) | 20 |
| Boston, MA (+variants) | 18 |
| Albany, NY | 8 |
| Nashville, TN | 7 |
| Washington, DC | 6 |
| Miami, FL | 6 |
| Syracuse, NY | 5 |
| Detroit, MI (+variants) | 7 |
| Tucson | 4 |
| Omaha, NE | 4 |
| Hattiesburg, MS | 4 |
| Denver, CO | 4 |
| …60 distinct origins total |
| Origin → Destination | Searches |
|---|---|
| Boston, MA → New York, NY | 9 |
| Albany, NY → Providence, RI | 4 |
| Washington, DC → New York, NY | 3 |
| New York, NY → Providence, RI | 3 |
| Nashville, TN → Chicago, IL | 3 |
| Nashville, TN → Atlanta, GA | 3 |
| Mount Vernon, WA → Chicago, IL | 3 |
| + 24 more pairs searched twice, 93 pairs searched once (123 distinct total) |
Across the Explore surface’s MCP search-trips calls, only one
(Chicago → Detroit, 5 trips) came back with results. Everything else returned empty.
This is the single most important thing to dig into — users are searching real routes and seeing nothing.
Every search.completed event was BOSSST → NYCPAS (Boston → NYC),
returned 0 trips with all 7 carriers failing (GLI, MGL, PPP, BLT, AMT, LML, TDS).
These ran only from dev/preview pods, consistent with the VPN/VPC requirement for pservs.
On 2026-06-01 (19:50–19:53) there were 3 wapi.auth_failed timeouts against
api.dev.wanderu.dev/v2/auth.json, which caused that day’s 4 autocomplete failures.
Nothing since. Plus 1 isolated llm.error (model didn’t call its tool).