Itinerary Quality Baseline

andaman-beach-couple BROKEN heavy 2.3/5

India from Chennai · 6d · kind ["beach"] · couple · budget premium · saves none

Catastrophic destination mismatch: couple asked for Andaman Islands beach trip, got Delhi-Goa-Jaipur city tour with one beach day

Itinerary realism 3 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 3  ||  Vibe/kind fit 1 · Companion fit 3 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Hotels are real, premium-tier and well-rated: ITC Grand Goa Luxury Collection (₹17,393/night, 4.7) in Arossim and Jai Mahal Palace Jaipur (₹14,978/night, 4.7) in Ganpati Nagar are appropriate luxury picks for a 'premium' couple
• Transport segments are sane modes/durations for the (wrong) cities chosen: Chennai->Delhi 160min flight ₹17,310, Delhi->Goa 145min, Goa->Jaipur 125min, Jaipur->Chennai 155min - no impossible drives, no 0-min legs
• Day 4 Jaipur sequencing is geographically tight and time-realistic: Hawa Mahal 07:00 -> Amber Fort 09:15 -> Albert Hall 12:15 -> Johari Bazaar 15:00 -> Dal Baati dinner 19:00 reads like a real Pink City day with proper meal anchors
• Goa Day 2 beach-shack rhythm is on-brand for the requested 'beach' kind: Lazy Dog Lounge breakfast -> Baga Beach swim -> Britto's lunch -> sunset shoot - real named venues, plausible durations

Bad

• CATASTROPHIC PLANNING FAILURE: user asked for Port Blair + Havelock (Andaman Islands) beach trip, flow produced Delhi(1n)+Goa(3n)+Jaipur(2n) - zero overlap with intended destination. A couple expecting island ferries, Radhanagar Beach, and scuba at Havelock will instead land in a North-Indian heritage circuit with a mainland beach stopover. This itinerary should never have shipped.
• Trip name 'Goa & Jaipur: 6 Days' silently omits the Delhi night entirely - the build-overview shows Delhi(1n) but no Day 0/1 Delhi activities exist in the itinerary days (Day 1 is 'Chennai to Goa'), so a paying user loses a night of planning
• Day 5 'Royal Hideaway' is geographically broken: Boutique Luxury Retreat Rajasthan at (26.568,72.346) is ~180km west of Jaipur near Pali/Jodhpur, yet day mixes it with 'Amer Fort Ramparts Walk' at (26.985,75.851) - that's ~280km away by road, both starting at 11:15/11:30 the same morning. Physically impossible.
• Day 5 contains a duplicate/conflicting visit: 'Amber Fort Elephant-Gate Ascent' already done on Day 4 (09:15, 135min, ₹550) reappears as 'Amer Fort Ramparts Walk' on Day 5 (11:30, 90min, ₹550) - same fort, same coords, charged twice
• Day 5 has two simultaneous activities: 'Countryside Walk & Village Trail' at 11:15 (120min) and 'Amer Fort Ramparts Walk' at 11:30 (90min) - 15-min overlap, two different locations 280km apart
• Day 6 is a sham 'last day': only Johari Bazaar (09:00, 90min) and Hawa Mahal (10:45, 45min, ALREADY visited at sunrise on Day 4) - no departure transport segment shown, no airport transfer, despite Jaipur->Chennai flight on the books
• Hawa Mahal is visited TWICE: Day 4 07:00 'at Sunrise' (90min, ₹100) and Day 6 10:45 (45min, ₹100) - duplicate filler stop, no narrative reason given
• Despite avoidance=['early_starts'], Day 4 schedules 'Hawa Mahal at Sunrise' at 07:00 - direct violation of the user's stated constraint
• Day 3 transit 'Goa -> Jaipur' is 270min for an actual 125min flight (per build-overview) and priced ₹7,000 vs the segment's ₹15,620 - inconsistent with the logistics layer
• Saved-place utilization is misleading: user's only Jaipur save is 'Boutique Luxury Retreat Rajasthan' which IS used Day 5, but at coords (26.568,72.346) it isn't actually in Jaipur - the place itself appears to be mis-geocoded or the wrong canonical venue was matched
• Trip kind=['beach'] gets just 2 days of actual beach (Day 1 partial + Day 2) out of 6 - the rest is heritage/palace/bazaar content that ignores the explicit single-kind intent

Issues: INTENT MISMATCH: user requested Andaman Islands; required Chennai->Port Blair flight + Port Blair->Havelock ferry never materialized because route collapsed to mainland; Day 3 transit duration ₹7,000/270min disagrees with build-overview Goa->Jaipur flight 125min/₹15,620 - itinerary day-card numbers don't reconcile with selected flight; Day 6 has no departure transport segment despite Jaipur->Chennai flight booked - user has no idea when to leave hotel; Day 5 places (Boutique Retreat near Pali ~180km west of Jaipur AND Amer Fort north of Jaipur) implies undocumented ~500km of intra-day driving with no transport legs scheduled; Per-person total ₹2,00,795 for a 6-day premium domestic trip is reasonable for the chosen cities, but the user is paying for 4 flights + Delhi night they didn't see itinerized; Activities total ₹50,250/pax is inflated by the ₹30,000 'Boutique Luxury Retreat Rajasthan' line which appears to be an overnight stay charged as a daytime activity (10:00, 60min) - either a duplicate hotel charge or mis-categorized; Day 4 + Day 6 both charge ₹100 for Hawa Mahal entry - duplicate visit means duplicate charge for the same monument

goa-beach-friends BROKEN heavy 3.3/5

India from Mumbai · 4d · kind ["beach"] · friends · budget mid · saves thin

Broken Day 1 chronology and a Day 4 prose-vs-data contradiction sink an otherwise sane Goa weekend

Itinerary realism 2 · Saved-place utilization 4 · Route & transport 3 · Budget & cost 4  ||  Vibe/kind fit 4 · Companion fit 4 · Pace fit 2 · Prose quality 3 · Day balance 2

Good

• Surfaces the user's only Goa save - 'Saree Store in Margao' is slotted at Day 3 11:00, paired sensibly with a 'Margao Old Market Walk' lead-in at 09:00 in the same locality (15.27-15.29 lat).
• Mumbai -> Goa overnight train at 600 min for ₹636 is a realistic, budget-appropriate mode pick for a 4-day mid-tier weekend (flight also valid; train is reasonable).
• Hotel pick 'Bloom Hotel - Dona Paula' at ₹1,842/night fits mid tier and a Panaji base; total ₹14,804 per person is plausible for a 4-day Goa weekend.
• Day 2 sequencing on the North Goa belt is geographically coherent: Anjuna (15.573,73.741) -> Brittos at Baga (15.554,73.751) -> Vagator (15.585,73.743) all within a tight cluster.
• Day 3 dinner 'Antares by Goan Chef Sarah Todd' is a real, well-known Vagator venue and fits the 'Coastal Flavors' theme.

Bad

• Day 1 chronology is BROKEN: 'Mumbai -> Goa Overnight Train' is listed at 20:00 (840 min duration) but is followed in the SAME day by 'Settle into Panaji Base' at 10:30 and 'Basilica of Bom Jesus' at 14:00 - the train departs AFTER the morning Panaji activities it precedes. A normal user would read this as nonsense.
• Day 4 internal contradiction: prose says 'Morning departure' but the only data row 'Goa to Mumbai Overnight Train' is at 23:30 (an evening/night departure, 480 min). Prose and data disagree.
• Day 4 is effectively empty for a 'packed' pace - one transit row at 23:30 and nothing during the day; a packed friends trip should have at least a beach morning + lunch before the evening train.
• Day 1 post-arrival is thin for packed pace: only 'Settle into Panaji Base' (120 min) + 'Basilica of Bom Jesus' (90 min) - no dinner anchor, no evening activity, no Fontainhas walk despite being Panaji-based.
• Day 3 has a 7.5-hour dead gap between 'Saree Store in Margao' ending ~12:30 and 'Antares' at 18:30 with no lunch or afternoon stop - meal anchors missing, breaks the packed pace promise.
• Hotel meta shows 'Murda' as the neighborhood for 'Bloom Hotel - Dona Paula' - Dona Paula is the neighborhood; 'Murda' looks like a geocode/neighborhood-resolver glitch (Murda is not a known Panaji locality).
• Day 1 'Basilica of Bom Jesus' coords (15.501,73.912) are in Old Goa, ~10km east of the Panaji base - fine, but no transit time is accounted for between 'Settle' at 10:30 (120 min, ends 12:30) and 14:00 Basilica - tight but acceptable; combined with the broken train-after-arrival ordering it compounds the realism hit.
• Activities total ₹6,800 across 4 days but Day 1 and Day 4 carry ₹0 activity cost - all spend is concentrated in Days 2-3 which is a thin distribution for a 'packed' trip.

Issues: Day 1: outbound 'Mumbai -> Goa Overnight Train' is listed AT 20:00 within Day 1 but the day also contains 10:30 and 14:00 Goa activities - the leg should be Day 0 evening -> Day 1 morning arrival, not Day 1 evening.; Day 4: 'Goa to Mumbai Overnight Train' at 23:30 contradicts the day prose 'Morning departure' - either the prose is wrong or the leg should be a Day 4 22-23h departure with arrival Day 5 morning (4-day trip framing is then ambiguous).; No intra-Goa transport modeled - North Goa (Anjuna/Vagator/Baga) to a Panaji base and to South Goa (Margao) are 25-35km hops each; trip implicitly assumes cab/scooter but never surfaces it.; Day 4 'Goa to Mumbai Overnight Train' carries cost=₹3,000 inside the activities/transit row even though the build-overview transport segment is ₹636 for Mumbai->Goa - return leg cost is inconsistent with the outbound and not reflected in the cost summary (flights=₹636 only).; Hotel 'Bloom Hotel - Dona Paula' at ₹1,842/night × ?n in logistics; cost summary hotels=₹7,368 implies 4 nights, but Day 1 is an overnight train (no hotel night used) and Day 4 ends on another overnight train - true hotel nights are likely 2-3, not 4. Hotel total is over-counted by 1-2 nights.

greece-luxury-couple BROKEN heavy 2.0/5

Greece from Mumbai · 8d · kind ["beach"] · couple · budget luxury · saves moderate

Geo-impossible 'trains' between Greek islands and a luxury trip with zero hotel cost - unusable as planned.

Itinerary realism 3 · Saved-place utilization 3 · Route & transport 1 · Budget & cost 1  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Uses 5 of the user's saved places by name: Plaka Neighborhood (Day 1), Ancient Agora of Athens (Day 3), Agios Nikolaos Harbour (Day 4), Blue Caves + Agios Sostis Harbour (Day 5), Myzithres Viewpoint + Xigia Beach + Plakaki Beach (Day 6), and Oia Village + Pyrgos Village (Days 7-8).
• Athens day themes are coherent and sensibly ordered: Acropolis/Parthenon/Erechtheion clustered Day 2, Agora + Temple of Hephaestus + Benaki Museum Day 3 (matching real geography).
• Day 8 Fira-to-Oia caldera hike (210min, 07:30 start) is a real, named, well-paced flagship Santorini activity.
• Avoidance of 'early starts' mostly respected on Athens days (08:00-09:00 starts are reasonable for a couple).

Bad

• GEO-IMPOSSIBLE: 'Athens -> Zakynthos | mode=train | 324 min | Rs 650' - there is no train to Zakynthos; the island is reached by ferry from Kyllini (or flight). Day 4 prose 'scenic train journey from Athens through the Peloponnese' is fabricated.
• GEO-IMPOSSIBLE: 'Zakynthos -> Santorini | mode=train | 186 min' (and Day 7 says '5.4-hour scenic train ride to Santorini') - Santorini is a Cyclades island with no train; this leg requires a ferry or flight (usually via Athens). Two of three intercity legs are physically impossible.
• LUXURY-TIER FAIL: only one hotel (Castelli Hotel, Zakynthos) is listed, with '-/night, ?n, rating=0' and hotels_total=0 in the cost summary. A luxury couple's 8-day trip with zero priced hotels and no Santorini/Athens stays is unusable - this is the structural [FAIL] check on hotels_priced.
• ROUTE ORDER nonsensical for an Athens arrival: Day 1 lands in Athens (Plaka), Days 2-3 are full Athens days, but the route summary is 'Santorini (3n) -> Zakynthos (3n) -> Athens (2n)' and Day 4 then 'trains' from Athens to Zakynthos. The header route contradicts the day-by-day, and Santorini gets only days 7-8 despite being listed as 3 nights.
• Mumbai -> Santorini flight at Rs 15,000 economy-rate with 240min duration is implausible: Mumbai-Santorini has no direct service, real luxury fares to Greece run multiples higher (Rs 60k-1L+), and 4h is far short of the real ~10-12h+layover routing.
• Day 6 repeats 'Xigia Beach exploration' which already appeared on Day 5 - duplicate stop.
• Day 4's only Zakynthos activity after a 5.5h fictional train is 75min at Agios Nikolaos Harbour - thin day.
• Saves missed: Pyrgos Village is used but Pisines Beach, Keri Caves, Cave of Love, Eros Cave are all unused despite Zakynthos having 11 saves and 3 nights allocated. Mykonos save (Negrita Mykonos) ignored - acceptable since Mykonos not in route.
• Day 8 ends at Pyrgos Village 11:30-12:45 with no return-flight transport segment shown on the final day - the Athens -> Mumbai flight is listed in segments but not anchored to a day; departure logistics unclear.

Issues: Athens -> Zakynthos labelled 'train' (324 min) - no rail service to Zakynthos; must be ferry from Kyllini or short flight.; Zakynthos -> Santorini labelled 'train' (186 min, cost Rs 16,740) - Santorini is a Cyclades island with no train link; requires ferry or flight (usually via Athens).; Day-by-day order (Athens first, Santorini last) contradicts the route summary (Santorini -> Zakynthos -> Athens).; Mumbai -> Santorini at 240min flight time is implausible (real routing is ~10-12h+ via Athens or European hub).; Final-day departure to Mumbai not anchored to Day 8 - logistics gap.; hotels_total = 0 despite a luxury tier and 8 nights - structural FAIL flagged in checks.; Only one hotel (Castelli Hotel) listed and it has '-/night, ?n, rating=0' - no Santorini or Athens stays at all.; Castelli Hotel is a mid-range Zakynthos property, not 5-star+; for a luxury couple the Zakynthos pick should be Olea All Suite, Lesante Blu, etc.; Mumbai -> Santorini Rs 15,000 fare is far below realistic luxury (or even economy) intercontinental pricing - per-person total Rs 48,599 for an 8-day luxury Greece trip is roughly 5-10x too low.; Activities total Rs 17,575 over 8 days is plausible but the absence of hotel cost makes the total meaningless for budgeting.

Structural fails: hotels priced (cost_summary.hotels_total > 0)

himachal-hills-solo BROKEN heavy 1.5/5

India from Delhi · 7d · kind ["hill-stations"] · solo · budget budget · saves none

Catastrophic destination swap: user asked for Manali+Kasol hill-stations, got Goa+Jaipur beach-and-palace at ~₹79K on a budget-tier solo brief.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 1  ||  Vibe/kind fit 1 · Companion fit 3 · Pace fit 3 · Prose quality 2 · Day balance 2

Good

• Day 4 Pink City sequence (Hawa Mahal 06:00 -> Jantar Mantar 08:30 -> Johari Bazaar 10:30 -> Old City walk 14:30) is geographically tight and chronologically clean within Jaipur's old city.
• Day 5 Amber Fort -> Jal Mahal -> Nahargarh Sunset is a sensible Jaipur loop with a real golden-hour anchor.
• Jaipur -> Delhi return by train (270 min, ₹353) is the right mode and a realistic Shatabdi-class price for a budget solo traveler.

Bad

• CRITICAL DESTINATION FAILURE: request explicitly named Manali + Kasol (hill-stations trip_kind), but the trip is Bangalore (1n) -> Goa (3n) -> Jaipur (3n). Zero Himachal content. A user who asked for hill-stations would reject this on sight.
• CRITICAL GATEWAY FAILURE: 'Delhi -> Bangalore | flight | 320 min | ₹13,520' as a leg-1 detour from a Delhi-origin domestic trip is geographically nonsensical and burns the entire budget envelope; Bangalore is then assigned 1 night with no day-1 activity. Looks like an India-picker collapse where the user's home Bangalore got injected as a routing hub.
• Day 1 has only 'Anjuna Beach & Flea Market' after a 06:00 flight + transfer - violates the user's 'early_starts' avoidance and leaves the day empty (one activity for the entire afternoon/evening).
• Goa -> Jaipur train at 1742 min (29h) is brutal for a solo budget traveler with a 3-night Jaipur slot - either flight (₹4,500 as shown on Day 3) or a re-route to skip Goa entirely is appropriate; the build shows BOTH a Goa->Jaipur train segment AND a Day 3 'Goa -> Jaipur Flight' inside the itinerary, contradicting itself.
• Day 6 'Boutique Luxury Retreat Rajasthan' at ₹30,000 + 'Farm-to-Table Dinner' ₹8,000 on a BUDGET-tier solo trip is a ~₹38,000 single-day blowout - the entire day reads luxury, not budget backpacker.
• Hotel block in build-overview shows 'Hanwant Nagar' (Jodhpur neighborhood) for the Jaipur hotel; coords (26.568, 72.346) plotted for the Boutique Retreat AND the Hotel Check-In are in Jodhpur, ~330km from Jaipur (26.92, 75.82). Stay city/coords mismatch.
• hotels_total = 0 in cost summary despite a selected hotel row - pricing pipeline failure; user sees activities ₹58,250 + flights ₹20,696 but no stay cost across 6 nights.
• 'Antares by Goan Chef Sarah Todd' priced at ₹2,500 solo + 'Local Spice Market Stop' at ₹3,000 + 'Johari Bazaar & Bapu Bazaar' at ₹5,000 + 'Bapu Bazaar' AGAIN on Day 7 at ₹3,000 (duplicate) inflate the activities line; Bapu Bazaar is double-billed across Day 4 and Day 7.
• 'Brittos Beach Shack' (Day 2 lunch ₹800) followed by 'Baga Beach Walk' at the same coords - filler stacking on top of a single beach strip.
• Day 7 ends at 11:45 with a kachori stop and Bapu Bazaar - no departure transport rendered in the day (Jaipur->Delhi train exists in segments but no anchor on the day card); leaves the day feeling truncated.

Issues: Delhi -> Bangalore flight (320 min, ₹13,520) as leg 1 of a Delhi-origin domestic trip is a routing hallucination; gateway logic appears to have used user's Bangalore home as a hub.; Goa -> Jaipur listed twice with contradictory modes: build-overview segment says train 1742 min ₹2,575; Day 3 itinerary says 'Goa -> Jaipur Flight' 315 min ₹4,500.; Bangalore -> Goa flight (75 min, ₹4,248) at the expense of skipping the intended Manali/Kasol road-from-Delhi route entirely.; No transport modeled to Manali/Kasol despite explicit request; correct mode would have been HRTC Volvo or shared taxi from Delhi/Chandigarh, ~12-14h.; hotels_total = 0 despite 6 paid nights across Goa + Jaipur - stay cost invisible to user.; Total ₹78,946 per person for a budget-tier solo 7-day domestic trip is roughly 2-3x what a backpacker brief implies (realistic budget Manali+Kasol 7d ex-Delhi: ₹15-25K).; Day 6 luxury retreat ₹30,000 + farm-to-table ₹8,000 = ₹38,000 single day flatly contradicts the budget tier.; Activities total ₹58,250 includes ₹5,000 + ₹3,000 + ₹3,000 in bazaar 'spend' line items which look like shopping budget padding rather than actual activity cost.

Structural fails: hotels priced (cost_summary.hotels_total > 0)

indonesia-bali-solo BROKEN heavy 2.3/5

Indonesia from Bengaluru · 9d · kind ["beach"] · solo · budget mid · saves moderate

Solo slow beach trip violates 'no early starts', adds an unrequested Jakarta urban detour, and breaks intra-day ordering on Nusa Penida

Itinerary realism 2 · Saved-place utilization 3 · Route & transport 2 · Budget & cost 2  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 1 · Prose quality 3 · Day balance 2

Good

• Uses Tegallalang Rice Terraces, Ubud Monkey Forest, Mount Batur Sunrise Trek, Quad Biking in Ubud, South Lombok Beaches, South Lombok Cafes and Bars, Lombok MotoGP Track, North Lombok and Lombok Island Overall - 8 of the user's saved place names appear
• Day 4 correctly uses a 'Bali -> Lombok Ferry' (285min, ~₹520) rather than a drive between islands
• Bali -> Lombok intra-archipelago hop also priced as a 40min flight (₹10,271) as an alternative segment - mode is at least island-aware
• Day-1 prose 'Flight from Bengaluru to Bali (3.5h), settle into Ubud, evening exploration' correctly recognises Ubud as the Bali landing zone

Bad

• Day 3 'Cliffs & Hidden Coves' is structurally broken: 'Nusa Penida Full Day Tour' starts at 08:00 for 540min, then 'Kelingking Beach' at 17:15, but 'Broken Beach' is scheduled at 12:00 and 'Angel's Billabong' at 14:00 - inside the full-day tour window - and 'Return to Ubud' is at 15:30, BEFORE Kelingking at 17:15. Times are non-monotonic and physically impossible
• User asked for Bali + Ubud + Lombok (beach, solo, slow) but the trip injected JAKARTA for 3 nights (Days 7-9) - a polluted, traffic-choked megacity with zero beaches. This is a serious intended-vs-actual cities miss and directly contradicts trip_kind=['beach']
• User explicitly avoided early_starts, yet Day 2 'Mount Batur Sunrise Trek' starts at 03:00, Day 5 at 07:00, Day 6 at 07:30, Day 7 at 08:00, Day 9 at 07:00 - the single hardest constraint is violated on 5 of 9 days
• Pace=slow but Day 8 in Jakarta stacks 5 activities (Kota Tua, Jakarta History Museum, Glodok Chinatown, Ancol Dreamland, Grand Indonesia Mall) ending at 21:30 - this is a packed itinerary, not slow
• Hotel 'Kabeh Jati Nusa Penida Villas & Restaurant' is on Nusa Penida island (neighborhood 'Suana' is on Nusa Penida), not on the Bali mainland where Days 1-3 are based - guest would need a ferry every day to reach Ubud activities. Geographic mislabel of the Bali stay
• Day 4 prose says 'Morning drive to coast, 4.5h ferry to Lombok' but the only listed activity is the ferry itself (08:00 start) - no drive segment, and only one activity after arrival
• No hotel at all in Lombok for the 3 nights (Days 4-6) - hotels array only covers Bali and Jakarta, leaving the middle of the trip un-lodged
• Day 3 includes 'Return to Ubud' at 15:30 (180min) but the very next day (Day 4) leaves Bali for Lombok - the user is staying in Bali but the implied lodging for night 3 is unclear given the Nusa Penida-island hotel
• Zero use of high-signal Canggu saves (Blue Point Beach, Finns Beach Club, Seminyak Beach) despite trip_kind=['beach'] and 3 saves there - Canggu is the obvious beach base in Bali and isn't visited
• Nusa Penida day appears but doesn't credit/use the user's two Nusa Penida saves ('Beautiful Beaches in Nusa Penida', 'Nusa Penida') by name
• Bali -> Lombok shows BOTH a 40min flight (₹10,271) in transport segments AND a ferry on Day 4 - duplicate/conflicting transport for the same leg
• Total per-pax ₹84,886 for 9d-international on mid-tier with no Lombok hotel and a ₹1,465/night Bali villa (sub-budget tier) feels under-priced; flights ₹59,857 then hotels only ₹13,851 for 6 priced nights is suspicious

Issues: Jakarta inserted into a beach-trip route with no user intent - adds 2 extra flights (Lombok->Jakarta ₹19,586 + Jakarta->Bengaluru ₹15,000) versus a clean Lombok->Bengaluru return; Bali->Lombok shown as both 40min flight (₹10,271 in segments) AND ~4.5h ferry on Day 4 - the trip can't be both modes on the same leg; Day 4 prose mentions 'Morning drive to coast' but no drive activity is scheduled before the ferry; Day 7 transit is labelled 'Lombok -> Jakarta' at 08:00 but the prose says 'Return to Bali, fly to Jakarta' - inconsistent routing description (direct vs via Bali); Bali hotel ₹1,465/night is sub-budget tier (mid would be ₹4-8k); also the property is on Nusa Penida island, not the Bali mainland the days reference; No Lombok hotel for 3 nights - missing line item; Cost summary hotels_total ₹13,851 covers only Bali (3n × ₹1,465 = ₹4,395) + Jakarta (3n × ₹3,152 = ₹9,456) = ₹13,851 confirming Lombok was simply skipped from lodging cost; Mount Batur Sunrise Trek at ₹350 is unrealistically low (real cost is ₹2,500-4,000 with guide+jeep+breakfast); Lombok->Jakarta flight at ₹1,300 in the day card vs ₹19,586 in the transport segment - 15x discrepancy between cost surfaces; Jakarta->Bengaluru in day card ₹1,800 vs ₹15,000 in transport segment - same large discrepancy

italy-classic-couple BROKEN heavy 2.0/5

Italy from Delhi · 11d · kind ["art-heritage"] · couple · budget premium · saves moderate

Classic triangle skeleton with grossly mispriced activities, a phantom Milan night, and a backwards Day 8 that breaks the route.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 2 · Budget & cost 2  ||  Vibe/kind fit 3 · Companion fit 3 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Day 2 Rome anchor stack is iconic and locally coherent: Colosseum 09:00, Roman Forum 11:15, Palatine Hill 13:00, Pantheon 15:30, Trastevere Dinner 19:30 - geographically tight and matches saved 'Colosseum'.
• Day 4 in Florence pulls four real saved places into one walkable arc: Uffizi Gallery, Spezierie Palazzo Vecchio, Slitti, Antica Erboristeria San Simone.
• Intercity legs are correctly TRAIN, not flight (Milan->Rome 180min, Rome->Florence 90min, Florence->Venice 120min, Florence->Rome 140min) - matches Italy reality of Frecciarossa/Italo.
• Day 6 Venice surfaces user's saved 'Burano' and 'Gondola Ride' as 'Gondola Through Grand Canal' and 'Burano Colorful Streets'.
• Avoidance honored: no activity before 08:30 on full days; respects 'early_starts' avoidance.

Bad

• ROUTE BREAK: Day 8 narrative says 'Return to Florence' but the segment is labeled 'Florence -> Venice Train' at 09:00 - directionally backwards and contradicts the actual route Venice(3n)->? The trip is supposed to fly home from Venice but the itinerary then spends Day 9-10 in Tuscany/Florence and Day 11 takes 'Florence -> Rome Train' to FCO, not Venice. The route logistics (Venice as final city, Venice->Delhi flight) directly contradict the day-by-day (ends in Florence, departs from FCO via Rome).
• PHANTOM MILAN NIGHT: Route says Milan 1n but Day 1 itinerary lands at Rome FCO at 15:00 and goes straight to Piazza Navona - Milan is never visited. The Milan hotel (Boutique Hotel Martini 17, ₹18,318) is paid for but unused; transport 'Milan -> Rome train 180min' has no corresponding day.
• COST UNIT BUG (severe): Activities are off by orders of magnitude. Uffizi Gallery ₹250 (real ~₹2,500), Spezierie ₹30, Slitti Gelato ₹8, Antica Erboristeria ₹25, Pantheon ₹9,600 (real entry ~₹2,000), Colosseum ₹34,560 (~17x real ~₹1,800), Venchi ₹29, Eataly ₹30, Burano ₹9. Most look like euros-as-rupees or paise-as-rupees; Colosseum/Pantheon look like inflated/wrong-currency. Activities total ₹60,765 is implausible vs the line items shown.
• HOTEL TIER MISMATCH: Premium tier in Italy should be 4-5* (₹35-60k/night minimum for Rome/Florence/Venice). Domus Appia 154 in Rome ₹11,684/n is a budget B&B in Appio-Latino, far from centro storico. B&B San Remigio Florence at ₹24,799 is mid-tier B&B not premium. Only Hotel Olimpia Venice ₹28,384 approaches mid - none are premium-tier Italy.
• SAVED PLACES BARELY USED: User saved 6 specific Rome places (Colosseum, Piazza Navona, Pantheon-adjacent, Vatican City, St. Peter's Basilica, Trastevere) - itinerary uses Colosseum, Piazza Navona, but MISSES Vatican City and St. Peter's Basilica entirely across 4 Rome nights, which is unforgivable for an 'art-heritage' premium trip. Saved 'Teatro La Fenice', 'St. Mark's Basilica', 'Piazza San Marco' in Venice all MISSED. Saved 'Galleria Vittorio Emanuele II' and 'Milan Duomo' missed (Milan unvisited).
• DUPLICATE/INVENTED DAY 8: 'Florence -> Venice Train' on Day 8 contradicts the narrative ('Return to Florence') AND is labeled cat=activity not transit, with cost ₹93 - the same row as Day 5's identical leg. Looks like a copy-paste bug.
• DAY 9 GEO-NONSENSE: Pienza (43.124,11.747) at 09:00 for 150min, then La Vecchia Bottega (43.077,11.680) at 12:30, then Conservo (43.777,11.253) at 15:30, then Eataly Firenze (43.825,11.141) at 17:00. Pienza is ~140km south of Florence; driving back to Florence Conservo at 15:30 after a 12:30 Pienza lunch is physically tight with no transit row. Val d'Orcia drive on Day 8 (210min at 14:00 from Venice???) is geographically impossible.
• DAY 3 EMPTY: Only train (10:30) + Arno walk (17:30) - 6 hour gap with no anchor; weak for a slow-pace premium trip.
• DAY 7 'Rest & Wander' has only 2 activities (Bacaro Crawl + Campo Santa Margherita) - reasonable for slow pace but a Venice premium day with no Doge's Palace / St. Mark's mention is a content miss given user's explicit saves.
• PROSE: 'Twilight dinner', 'island light', 'Final Florence wanderings' are competent but generic; no editorial specificity. 'Trattoria Dinner' as an activity name is placeholder-tier.

Issues: Milan night booked + Milan->Rome train segment exists, but Day 1 itinerary lands at Rome FCO directly - phantom city with paid hotel never visited.; Day 8 labeled 'Florence -> Venice Train' while narrative is 'Return to Florence' - direction reversed AND duplicates Day 5's identical leg.; Day 11 'Florence -> Rome Train' to FCO contradicts route logistics showing Venice->Delhi as the departure flight; the day-by-day departs from a different city than the booked international flight.; Day 8 places Val d'Orcia scenic drive at 14:00 with no transit row from Venice/Florence - geographically impossible without a 3-4h transfer.; Day 9 jumps Pienza (43.124) -> La Vecchia Bottega (43.077) -> Conservo Florence (43.777) -> Eataly Florence (43.825) in afternoon with zero transit rows between them - ~140km moves un-modeled.; Colosseum ₹34,560 for 120min: real entry ~€18 (~₹1,700) even with Arena access (~€24); inflated ~17-20x or wrong unit.; Pantheon ₹9,600: real entry €5 (~₹470); inflated ~20x.; Uffizi Gallery ₹250: real ticket €25-30 (~₹2,500-3,000); 10x UNDER - looks like euros stored as rupees.; Slitti Gelato ₹8, Burano ₹9, Murano transit ₹0, Spezierie ₹30, Venchi ₹29, Eataly ₹30, Conservo ₹30, La Vecchia Bottega ₹30 - all look like euros stored as rupees (1 EUR ≈ ₹93, not ₹1).; Train fares: Rome->Florence ₹347 (~€3.70), Florence->Venice ₹306 (~€3.30), Milan->Rome ₹716 (~€7.70) - Frecciarossa real fares are €40-90 booked ahead; these are ~10x UNDER, again EUR-as-INR suspect.; Premium-tier hotels: Domus Appia 154 (₹11,684) is 2-3* B&B; B&B San Remigio (₹24,799) is mid-tier; only Venice Olimpia (₹28,384) approaches the range. Premium Italy expectation is €350-600/n (~₹32-55k); none of the picks match the requested tier.; Activities subtotal ₹60,765 doesn't reconcile with line items (sum of shown activity costs ≈ ₹59,000 but mix of inflated landmarks + under-priced shops suggests two different cost-unit pipelines colliding).; Total per-person ₹3,16,737 for 11d premium Italy is LOW (realistic premium baseline ~₹5-7L per pax for Rome+Florence+Venice with 4* hotels + Frecciarossa) - confirms hotels are under-tier and activities under-priced.

italy-scenic-friends BROKEN heavy 1.8/5

Italy from Bengaluru · 8d · kind ["cities"] · friends · budget mid · saves rich

Generic Rome+Florence trip that ignores the user's scenic-Italy saves (Dolomites 19, Positano 15, Cinque Terre 9, most of Lake Como 21) and violates the early_starts avoidance.

Itinerary realism 2 · Saved-place utilization 1 · Route & transport 2 · Budget & cost 2  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 2 · Prose quality 3 · Day balance 2

Good

• Villa del Balbianello on Day 2 is a real, high-signal Lake Como save and is correctly placed near Como.
• Florence cluster on Day 4 (Slitti, Officina Profumo-Farmaceutica di Santa Maria Novella, Spezierie Palazzo Vecchio, Venchi, Eataly Firenze) genuinely lifts five of the seven user Florence saves into a single walkable artisan day.
• Train mode is correctly used for the intra-Italy legs (Lake Como -> ... -> Rome 90min Florence -> Rome) instead of forcing a domestic flight.

Bad

• Saved-place utilization is catastrophic for a 'rich' saves tier: 0 of 19 Dolomites saves used (no Lago di Braies, Seceda, Tre Cime, Val di Funes), 0 of 15 Positano saves (no Positano, Le Sirenuse, Fornillo Beach), 0 of 10 Amalfi Coast saves, 0 of 9 Cinque Terre saves (no Manarola, Riomaggiore, Vernazza), 0 of 9 Capri saves. Lake Como got only 1 night despite being the top save (21).
• Intended cities were Lake Como + Dolomites + Positano; flow shipped Lake Como + Rome + Florence. Two of the three user-intended destinations are simply missing from the trip.
• Route geometry is inefficient: ending in Rome but returning to Florence for the Day 8 departure ('Train to Florence airport, fly home') wastes a half day and ignores Rome FCO as the obvious intl hub when the user lands in Milan and the last 3 nights are in Rome.
• Day-order vs logistics mismatch: build-overview lists Lake Como -> Rome -> Florence (4n / 3n) but the itinerary actually runs Como (D1-2) -> Florence (D3-4) -> Rome (D5-7), with Day 3 transit named 'Lake Como -> Florence' and Day 5 transit 'Florence -> Rome'. The two artifacts disagree on the route.
• Violates the user's stated 'early_starts' avoidance repeatedly: Colosseum 08:00 (Day 6), Vatican Museums 08:30 (Day 7), Day 8 transit 07:30.
• Activity prices are wildly inflated: Colosseum and Roman Forum at ₹17,280 each (~€185 — actual entry ~€16/₹1,500), Florence Duomo climb ₹18,600 (~€200, real ~€30), Uffizi ₹24,960 (~€267, real ~€26), Vatican Museums ₹27,900, St. Peter's ₹18,600 (basilica entry is free). Activities subtotal ₹2,11,834 is ~5-10x reality and dominates the trip cost.
• Greenway del Lago coords (45.525, 9.171) sit in Milan, not on the Lake Como Greenway (which is north of Como ~45.99,9.24). The card mislocates a real saved place and prices a free public walking path at ₹15,500.
• Hotels persisted only one row (Edition Lake Como) with no nightly rate, no nights count, and hotels_total=0 — structural check flagged this. Rome (4n) and Florence (3n) have no hotel at all.
• Day 5 is essentially a half-day write-off: only 'Trastevere Evening Walk' after a 09:00->10:30 train, no lunch, no afternoon. Day 8 has zero activities, only the airport transit.
• Day 1 misses obvious user saves: Bellagio, Varenna (Vecchia Varenna gets a dinner mention on D2 but the town itself isn't visited), Villa Carlotta, Villa Monastero, Villa Melzi, Funicular to Brunate — 1n in Como with 21 saves is dramatically under-served.

Issues: Inconsistency between build-overview route (Lake Como -> Rome -> Florence) and itinerary day order (Como -> Florence -> Rome): the two artifacts disagree on the city sequence.; Day 8 departs from Florence airport even though the last 3 nights are in Rome — backtracking ~280km when Rome FCO is the obvious hub.; Lake Como -> Rome listed as a 390min direct train; real Como->Rome requires a Milano Centrale change and runs ~4h15m. Plausible but the 'direct scenic train' framing on Day 3 ('4-hour scenic train to Florence') is misleading.; Greenway del Lago listed at Milan coordinates (45.525, 9.171) rather than the actual Greenway north of Como (~45.99, 9.24).; Structural check FAIL: hotels_total=0 — only Edition Lake Como persisted, no rate, no nights count, no hotels for Rome (4n) or Florence (3n).; Colosseum ₹17,280 and Roman Forum ₹17,280 vs real entry ~€16 / ₹1,500 — over 10x inflated.; Vatican Museums ₹27,900, St. Peter's Basilica ₹18,600 (basilica is free entry), Florence Duomo Climb ₹18,600, Uffizi ₹24,960 — all 5-10x real ticket prices.; Greenway del Lago priced at ₹15,500 for a free public lakefront walking path.; Activities subtotal ₹2,11,834 is implausibly the dominant line item; mid-tier Italy reality for sightseeing is closer to ₹15-25k per person for a museum-heavy 8 days.; Total per person ₹2,42,966 happens to land in a reasonable mid-tier zone only because hotels are zeroed out — a real mid-tier 7-night hotel stack would add ₹70k-1.4L, pushing realistic total to ₹3.0-3.8L.

Structural fails: hotels priced (cost_summary.hotels_total > 0)

japan-classic-couple BROKEN heavy 2.8/5

Japan from Delhi · 10d · kind ["cities"] · couple · budget premium · saves rich

Sane Tokyo-Kyoto-Osaka spine with correct Shinkansen routing, but the Osaka leg is a duplicate-stop loop (Dotonbori x3, Umeda Sky x2, Tsutenkaku x2, Tombori Cruise x2) and Day 8 places 'teamLab Borderless' at Tokyo coordinates while the user is in Osaka - a geo-impossible hallucination.

Itinerary realism 2 · Saved-place utilization 3 · Route & transport 4 · Budget & cost 2  ||  Vibe/kind fit 4 · Companion fit 4 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Intra-Japan transport modes correct: Tokyo->Kyoto and Kyoto->Osaka both flagged as train (138min and 18min) with no spurious intra-country flights.
• Saved Kyoto canon all hit on Day 4: 'Fushimi Inari Taisha', 'Kinkaku-ji (Golden Pavilion)', and 'Arashiyama Bamboo Grove' all match user saves verbatim.
• Day 7 'Nara Deer Park & Todai-ji' matches user's 'Nara Day Trip' / 'Nara Park' saves and uses a correct ~45min Osaka<->Nara train hop with a return leg.
• Osaka anchors pulled from saves: 'Dotonbori', 'Tombori River Cruise', 'Umeda Sky Building', 'Shinsekai Street Food Tour', 'Tsutenkaku Tower' all map to the user's 20 Osaka saves.
• Premium hotel selections sensible for couple+premium tier: Hanakohro Tokyo (4.7, ₹61,692/n), The Thousand Kyoto (4.5, ₹47,111/n), Conrad Osaka (4.6, ₹50,768/n) - all real, well-rated, neighborhood-appropriate.
• Route distribution Tokyo (4n) / Kyoto (3n) / Osaka (3n) is the textbook Golden Route split for a 10-day couples trip; gateway logic correct (arrive Tokyo, depart Kansai).

Bad

• GEO-IMPOSSIBLE: Day 8 ('Osaka Street & Art') opens with 'teamLab Borderless' at coordinates (35.660, 139.731) - that is Tokyo Odaiba/Toyosu, ~400km/2.3h Shinkansen from Osaka. The same hallucinated entry already appears on Day 2 in Tokyo, so the model double-booked the Tokyo attraction into an Osaka day.
• Massive duplicate-stop loop on the Osaka leg: 'Dotonbori' appears Day 5 ('Dotonbori Street Exploration'), Day 8 ('Dotonbori Street Walk'), AND Day 9 ('Dotonbori Final Night') - three visits in five days. 'Tombori River Cruise' repeats Day 5 + Day 8. 'Umeda Sky Building' repeats Day 6 (₹1,536) + Day 9 (₹876). 'Tsutenkaku Tower' repeats Day 6 ('Tsutenkaku Tower Walk & Dive') + Day 9 ('Tsutenkaku Tower'). Pure filler.
• User's 'early_starts' avoidance is violated twice: Day 4 'Fushimi Inari Taisha' starts at 07:00, Day 10 'Kansai Airport Transfer' starts at 06:30. Avoidances should be hard constraints.
• Headline Osaka saves are missing despite 20 saves and 3 Osaka nights: 'Osaka Castle', 'Universal Studios Japan' (an obvious couples pick), 'Rikuro Ojisan's Cheesecake - Namba Main Store', 'Osaka Amazing Pass' all skipped. With the duplicate-loop above there was clearly room.
• Headline Tokyo saves missing across 4 Tokyo days: only 'Asakusa & Senso-ji Temple' from the 58-save Tokyo list. 'Hamarikyu Gardens', 'Godzilla Head in Shinjuku', 'Kyu-Furukawa Gardens', 'Kodai-ji' all ignored. Day 2's 'Akihabara Electric Town', 'Shibuya Crossing & Hachiko', 'Conveyor Belt Sushi - Genki Sushi' are generic tourist picks, not from saves.
• Headline Kyoto saves missing despite 50 saves and 3 days: 'Gion District', 'Kiyomizu-dera Temple', 'Kichikichi Omurice', 'Byodo-in', 'Ine Fishing Village', 'Adashino Nenbutsu-ji' all skipped. Day 3 padded with generic 'Nishiki Market' and 'Pontocho Alley' instead.
• Absurd transport pricing: Tokyo->Kyoto Shinkansen at ₹540 (segment) / ₹572 (activity). Real Nozomi reserved is ~JPY14,000-22,000 = ~₹8,000-13,000 per person. Kyoto->Osaka at ₹65 - real is ~₹350-500. Off by ~15-20x and will mislead the user on budget.
• Multiple activity costs look like extraction failures: 'Kinkaku-ji' ₹4, 'Ryoan-ji Temple' ₹5, 'Nishiki Market' ₹34, 'Nara Deer Park & Todai-ji' ₹342, 'Return to central Kyoto' ₹6, 'Return to Osaka' ₹34. Real entry fees and JR fares are ~₹250-450 each. Almost certainly a paise/rupee unit bug.
• Day 1 timing is internally broken and inconsistent with the build-overview: build-overview lists Delhi->Tokyo flight at 240min (4h), Day 1 lists it at 360min (6h), and real DEL->NRT/HND is ~7h45m of flight + 3.5h timezone gain. Hardcoded 00:00 departure + 07:30 'airport transfer' implies a ~06:00 landing that no real flight actually achieves.
• Day 8 prose 'Osaka Street & Art' is undermined by Tokyo's teamLab being the anchor - the model literally couldn't find an Osaka art experience and lifted the Tokyo one.
• Day 9 'Last Night Osaka' is structurally vapid: Tsutenkaku 10:00 (already done Day 6), Umeda 16:00 (already done Day 6), Dotonbori 18:00 (already done Days 5 & 8). Zero new content for the trip-closing day.
• Day 3 'Tokyo -> Kyoto Shinkansen' has coordinates (35.081, 139.767) - that is Yokohama/Miura peninsula, not Tokyo Station nor Kyoto. Activity lat/lng anchor is wrong.
• Activity total ₹31,557 pp across 10 days is unrealistic for premium Japan - typical entry/experience spend for this trip shape is ~₹80K-₹1.2L pp. Combined with bottom-of-barrel train prices, total ₹6,02,567 will under-estimate the actual trip by a meaningful margin.

Issues: Day 8 'teamLab Borderless' placed in the Osaka block but at Tokyo coordinates (35.660, 139.731) - ~400km out of city, geo-impossible within an Osaka day.; Tokyo->Kyoto Shinkansen priced at ₹540/₹572 per person - real Nozomi reserved is ~₹8,000-13,000. Off by ~15-20x.; Kyoto->Osaka train priced at ₹65 in segment vs ₹570 in activity (inconsistent between layers); real is ~₹350-500.; Return Osaka->Nara segment costs ₹34 - real JR fare is ~₹450.; Day 3 Shinkansen activity coordinates (35.081, 139.767) point at Yokohama/Miura, not Tokyo Station or Kyoto Station.; Delhi->Tokyo flight scheduled at 00:00 for 240min (build-overview) vs 360min (Day 1) - both understate the real ~7h45m direct flight and disagree with each other.; Activity entry costs systematically under-reported by ~20-100x: Kinkaku-ji ₹4 (real ~₹250), Ryoan-ji ₹5 (real ~₹350), Nishiki Market ₹34, Nara Park ₹342 (Todai-ji entry alone is ~₹360). Looks like a paise/rupee unit bug.; Train costs as above - Shinkansen showing ~₹540 instead of ~₹10,000, Kyoto->Osaka ₹65 instead of ~₹450.; Activities total ₹31,557 pp for 10 days in premium Japan is unrealistic - real spend would be ~₹80K-₹1.2L pp.; Hotel pricing is the only realistic component (₹47K-₹61K/night for 4.5-4.7 rated hotels in Tokyo/Kyoto/Osaka is sane premium-tier).; Flight ₹15,000 each way DEL<->Japan in Oct is on the low side of plausible for premium-tier couples - typical economy is ₹40-70K return, premium economy ₹1.2-2L.

japan-foodie-solo BROKEN heavy 1.5/5

Japan from Mumbai · 7d · kind ["food"] · solo · budget mid · saves rich

Generic Tokyo+Hakone template ignores 63 real Japan saves, mis-geolocates 'Hakone' days to Mt Fuji/Kawaguchiko, routes lunch to Kyoto mid-Tokyo-day, and prices a 7-day Japan trip at ₹36k with zero hotels.

Itinerary realism 2 · Saved-place utilization 1 · Route & transport 2 · Budget & cost 1  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 2 · Prose quality 3 · Day balance 2

Good

• Day 2 'Senso-ji Temple - Asakusa' is a real near-match for the user's saved 'Asakusa & Senso-ji Temple' (Tokyo) - one of the only saved-place hits in the whole 7 days.
• Route shape Tokyo (4n) -> Hakone (3n) is a reasonable shell for the intended Tokyo+Hakone+Kawaguchiko trio, even though the third intended city (Kawaguchiko) was dropped.
• Day themes prose is punchy and on-brand ('Electric Tokyo', 'Art & Innovation', 'Mountain & Peaks') - no obvious AI filler-isms.
• Mumbai -> Tokyo as a flight (not drive) and no degenerate 0-min segments - basic transport sanity holds at the leg level.

Bad

• Saved-place utilization is essentially zero: user has 58 Tokyo + 5 Hakone + 4 Kawaguchiko + 50 Kyoto saves (incl. Hamarikyu Gardens, Kyu-Furukawa Gardens, Godzilla Head in Shinjuku, Meiji-related, Kichikichi Omurice, Ginza Nair's, and dozens of Indian restaurants flagged by a foodie - the explicit trip_kind), and the itinerary picks generic guidebook stops (Shibuya Crossing, Akihabara, teamLab, Ichiran, Genki Sushi, Tsukiji) instead. ~1 of 7 itinerary days touches a saved place.
• Day 6 'Vegan Ramen UZU Kyoto' has coords (35.015, 135.768) - that is Kyoto, ~450km from Tokyo. The day prose says 'Train back to Tokyo, evening exploration'. Geo-impossible as scheduled (14:00 lunch in Kyoto, 16:00 Meiji Shrine in Tokyo).
• Days 4-5 are labeled the 'Hakone' leg but the actual coordinates land at Chureito Pagoda (35.402, 138.779), Mt Fuji Fifth Station (35.361, 138.727), and Mt Fuji Sunrise Trek (35.362, 138.688) - all Fujiyoshida/Kawaguchiko/Mt Fuji, NOT Hakone (~50km away). Only 'Hakone Hot Springs & Lake Ashi' (35.233, 139.107) is actually in Hakone. The flow silently substituted Mt Fuji content for the dropped Kawaguchiko city.
• Day 5 'Mt. Fuji Sunrise Trek' starts at 07:30 - directly violates the user's stated avoidance: ['early_starts']. Sunrise treks are by definition pre-dawn (this is also off-season for the trek in March).
• Day 7 'Senso-ji Temple - Asakusa' (09:45) is a literal duplicate of Day 2's Senso-ji visit (08:00) - same coords (35.715, 139.797), same name. Lazy filler for a departure day.
• Meiji Shrine coords (35.276, 139.699) and Shinjuku Gyoen coords (35.284, 139.711) are both wrong by ~0.4 degrees latitude - real Meiji Shrine is ~35.676 N. Those coords drop both stops south of Yokohama, not in central Tokyo. Either the rows are mis-geocoded or the names are placeholders.
• Cost numbers are catastrophically broken across the board: teamLab Borderless ₹38 (real ₹2,400+), Ichiran Ramen ₹15, Genki Sushi ₹20, Tokyo->Hakone Shinkansen ₹57 (real Romancecar/Odakyu ~₹1,800 / Shinkansen ~₹2,500), Mt Fuji Fifth Station ₹29, Lake Ashi ₹29, Vegan Bistro Jangara ₹29. These look like a paise/rupee unit bug or unconverted cents. Activities total ₹6,116 for 7 days in Japan is fantasy.
• Mumbai <-> Tokyo flights at ₹15,000 each way is far below realistic (typical ₹40-65k one-way mid-tier). Total trip cost ₹36,256 with ZERO hotels persisted for a 7-day Japan trip is unusable - the user would have nowhere to sleep and would massively underestimate the real spend.
• Trip kind is 'food' but the itinerary has only 3 restaurants in 7 days (Genki Sushi, Ichiran, Vegan Bistro Jangara, plus the mis-located Vegan Ramen UZU Kyoto and Tsukiji market), and uses none of the user's many saved restaurants (Ginza Nair's, Kichikichi Omurice, Erick South Masala Diner, Andhra Dining, the Osaka Dotonbori/Rikuro Ojisan's cluster, etc.). This is not a foodie itinerary.
• Pace = 'packed' but most days carry just 2-3 activities (Day 1: 1 stop, Day 3: 2 stops, Day 7: 2 stops). A packed solo foodie would expect 4-6 anchors/day.
• Tokyo->Hakone marked as mode=drive (84 min, ₹140) - solo traveler from Mumbai isn't renting a car for a one-way mountain hop; this should be the Odakyu Romancecar or Shinkansen+Hakone-Tozan, and the price is off by 1-2 orders of magnitude.
• Intended cities Tokyo/Hakone/Kawaguchiko collapsed to Tokyo/Hakone - Kawaguchiko silently dropped despite 4 user saves there (incl. 'Mount Fuji Day Trip', 'Shiraito Falls', 'Hitsujiyama Park').

Issues: Day 6 routes user from Hakone -> Tokyo (Shinjuku) at 09:00, then to 'Vegan Ramen UZU Kyoto' (Kyoto coords) at 14:00, then back to Meiji Shrine Tokyo at 16:00 - geo-impossible without 2x Shinkansen Tokyo<->Kyoto same afternoon (~2h15 each way).; Tokyo -> Hakone marked as 'drive' (84 min, ₹140). Solo international traveler context: should be Odakyu Romancecar (~90min, ~₹1,800) or Shinkansen Kodama + Hakone-Tozan, not driving. Also intra-Japan short hop correctly NOT a flight - that part is right.; Days 4-5 advertised as 'Hakone' but coords place 4 of 5 stops at Mt Fuji / Fujiyoshida (~50km from the Hakone hotel base). Either the route needs a 3rd city (Kawaguchiko, as the user intended) or these stops need to be replaced with actual Hakone content (Owakudani, Hakone Open Air Museum, Lake Ashi pirate ship, Hakone Shrine).; Return leg Hakone -> Mumbai as direct flight at ₹15,000 ignores that the user must first travel Hakone -> HND/NRT (Shinkansen + Narita Express, ~3h) - day 7 has no transit time blocked for airport.; No hotels persisted (selected_hotels=0) - user has no lodging across 7 nights.; Activity costs look like a paise/rupee unit bug or unconverted cents: teamLab Borderless ₹38 (real ₹2,400+ / JPY 3,800), Ichiran Ramen ₹15 (real ₹700-900), Genki Sushi ₹20, Mt Fuji 5th Station ₹29, Lake Ashi ₹29, Vegan Bistro Jangara ₹29. These are 1-2 orders of magnitude too low.; Tokyo -> Hakone train priced at ₹57 - real Romancecar ~₹1,800, real Shinkansen+Hakone-Tozan ~₹2,500. Looks like the same unit bug.; Mumbai <-> Tokyo flights at ₹15,000 each way is well below realistic mid-tier (₹40-65k one-way is typical economy on the route, often higher in March).; Hotels total = ₹0 (selected_hotels=0). 7 nights Japan at mid budget realistically ₹1-1.5L total. Reported per-person total ₹36,256 understates the real trip by ~80%.; Cost ceiling sanity: ₹500 for Shinjuku Gyoen Park is also off (real entry ₹300 / JPY 500). ₹2,400 for Vegan Ramen UZU and ₹3,000 for Tsukiji Outer Market are the only two costs in the right order of magnitude - everything else looks broken at the unit-conversion layer.

Structural fails: hotels present; hotels priced (cost_summary.hotels_total > 0)

karnataka-nature-couple BROKEN heavy 1.5/5

India from Mumbai · 5d · kind ["heritage"] · couple · budget mid · saves moderate

Asked for Bengaluru+Coorg heritage trip, got a Delhi-Goa-Jaipur beach/palace mashup that ignored all 10 Bengaluru saves and forces a 29-hour Goa-Jaipur train.

Itinerary realism 2 · Saved-place utilization 1 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 3 · Prose quality 2 · Day balance 2

Good

• Hotels are real and rated (Bloom Hotel - Dona Paula 4.8 in Goa, Hotel Grand Lotus Inn 4.7 in Jaipur) with sane per-night pricing for the mid tier (₹1,842 and ₹1,293).
• Day 4 palace cluster in Jaipur (Jantar Mantar 09:30 -> Johari/Bapu Bazaar 11:30 -> Boutique Luxury Retreat) is geographically coherent within the Pink City core.
• Avoidance of early starts is mostly honored (Day 1 starts 14:30, Day 2 10:00, Day 3 09:00) - though Day 5 Hawa Mahal at 06:15 violates this directly.

Bad

• INTENDED-VS-ACTUAL CITY FAILURE: user asked for Bengaluru + Coorg (Karnataka heritage), flow produced Delhi (1n) -> Goa (2n) -> Jaipur (2n). Zero overlap with intended destinations - this is a planning collapse, not a trip.
• SAVED PLACES IGNORED: Bengaluru has 10 curated saves (Bar Spirit Forward, Copitas, Soka, Uno Izakaya, Muro, etc.) - none surface anywhere because Bengaluru is not in the route. Jaipur has 1 save (Boutique Luxury Retreat Rajasthan) which IS used Day 4 - the only save utilized out of 31 Indian saves.
• GEO-IMPOSSIBLE TRANSPORT: Goa -> Jaipur shown as a 1742-minute (29-hour) train at ₹2,575 on Day 3, while the itinerary day-3 prose says 'Morning departure from Goa. 4.5h flight to Jaipur' and lists a 'Goa -> Jaipur Flight' at 09:00 for 330min ₹8,000. The build-overview transport row contradicts the day itinerary - one of them is wrong, both can't be true.
• GATEWAY VIA DELHI IS NONSENSICAL: Mumbai -> Delhi -> Goa adds a 135-min flight (₹11,100) plus 1 wasted night in Delhi before Goa, when Mumbai -> Goa is a direct ~1h hop. Delhi is logged as a route city but has zero itinerary days assigned to it (itinerary jumps Mumbai -> Goa on Day 1).
• BUDGET MISMATCH: 'mid' tier total ₹1,01,575 per person with flights ₹37,605 and activities ₹57,700 is high - and a single Boutique Luxury Retreat at ₹30,000 plus City Palace at ₹7,500 and Jantar Mantar at ₹5,000 are 4-10x typical entry/stay costs (City Palace entry is ~₹700, Jantar Mantar ~₹200).
• HERITAGE VIBE PARTIALLY MET BY ACCIDENT: trip_kind=['heritage'] is served by Jaipur palaces but undermined by Day 1-2 Goa beach content (Anjuna Beach, Baga & Calangute, Brittos Beach Shack) which is leisure, not heritage.
• PROSE IS FILLER: 'Settle into beach vibes by evening', 'Golden-hour exploration, architectural photography, luxury retreat' read as AI-generic copy with no specificity to Karnataka heritage or the user's request.
• DAY 5 IS A STUB: single 06:15 Hawa Mahal stop (violates early_starts avoidance) then 'Departure transfer to airport' with no actual airport transit logged - half-day at best, billed as a full day.

Issues: Goa -> Jaipur listed as both a 1742-min train (build-overview) AND a 330-min flight (Day 3 itinerary) - contradictory; Mumbai -> Delhi -> Goa gateway adds an unnecessary 135-min flight + 1 Delhi night when Mumbai -> Goa is a direct ~1h flight; Delhi appears in route with 1 night but has zero itinerary days assigned (Day 1 jumps Mumbai -> Goa directly); Intended cities Bengaluru and Coorg appear nowhere in transport segments; Bengaluru -> Coorg should have been a ~5h drive, not flagged at all; City Palace Jaipur listed at ₹7,500 - actual entry is ~₹700 (10x inflation); Jantar Mantar listed at ₹5,000 - actual entry is ~₹200 (25x inflation); Boutique Luxury Retreat at ₹30,000 for a single 390-min stop is luxury-tier pricing on a 'mid' budget trip; Per-person total ₹1,01,575 is luxury-band for a 5-day domestic trip; mid-tier benchmark would be ₹40-60K

kerala-nature-couple BROKEN heavy 1.8/5

India from Bengaluru · 7d · kind ["heritage"] · couple · budget mid · saves none

Total destination substitution: user asked for Kerala monsoon (Kochi/Munnar/Alleppey) and got Goa+Jaipur with a 29-hour train leg, duplicate Hawa Mahal visits, and early-start violations the user explicitly avoided.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 1 · Companion fit 3 · Pace fit 2 · Prose quality 2 · Day balance 2

Good

• Day 4 Jaipur old-city block is geographically coherent: Hawa Mahal (26.924,75.827) -> Jantar Mantar (26.925,75.824) -> Johari/Bapu Bazaar -> Govind Dev Ji Aarti are all real, walkable, and well-ordered.
• Amber Fort Elephant-Gate Ascent (26.985,75.851), Jal Mahal (26.984,75.868), and Amber Village on Day 6 are real canonical Jaipur places with reasonable durations and a sensible sequence.
• Hotel pricing reads mid-tier and not luxury-inflated: Bloom Hotel Dona Paula ₹1,842/night and Hotel Grand Lotus Inn ₹1,293/night both plausible.
• It surfaced the two relevant saved places the user actually had in-route ('Saree Store in Margao' Day 2, 'Boutique Luxury Retreat Rajasthan' Day 5) rather than inventing pure junk.

Bad

• CATASTROPHIC SCOPE MISS: user asked for Kochi/Munnar/Alleppey (a classic Kerala nature loop in monsoon) and the trip produced Goa + Jaipur - zero Kerala, no Munnar tea estates, no Alleppey backwaters, no Fort Kochi heritage.
• Trip kind ['heritage'] is partially betrayed: Day 1-2 Goa (Anjuna Beach & Flea Market, Brittos Beach Shack, beach-shack seafood) is a beach trip, not heritage; only the Jaipur half nominally fits the requested vibe.
• Transport is internally contradictory: logistics shows Goa -> Jaipur mode=train, 1742 min (~29h) at ₹2,575, but Day 3 prose says 'afternoon flight to Jaipur (4.5h)' and the day card lists 'Goa → Jaipur Flight' 330min. The two sources disagree on mode entirely.
• If the realized leg is the 29h train, it is operationally insane for a 7-day mid-tier slow-pace couple's trip and burns all of Day 3 in transit.
• Avoidance ['early_starts'] is violated three times: Day 4 'Hawa Mahal at Sunrise' at 06:00, Day 5 'Nahargarh Fort Trek' at 08:00, Day 7 'Hawa Mahal at Sunrise' at 07:30 - the very thing the user said to avoid.
• Duplicate stops within a 7-day trip: 'Hawa Mahal at Sunrise' appears on BOTH Day 4 and Day 7, and 'Johari Bazaar & Bapu Bazaar' appears on BOTH Day 4 and Day 7 - filler, not curation.
• Day 5 is geographically impossible: 'Boutique Luxury Retreat Rajasthan' at (26.568,72.346) is ~350km west of Jaipur near Jodhpur/Pali, yet scheduled noon-5pm sandwiched between 'Nahargarh Fort Trek' 08:00-10:30 and 'Nahargarh Fort Sunset & Lights' 17:30-19:00 back in Jaipur. The round-trip cannot happen in those hours.
• Monsoon (12-18 July) is never acknowledged in any day prose; July is peak monsoon in Goa when beaches are flagged dangerous and Anjuna Flea Market is largely shut - 'evening beach walk' and 'Anjuna Beach & Flea Market' on Day 1 are seasonally inappropriate. Rajasthan in July is hot-monsoon too - sunrise treks at Nahargarh are weather-risky.
• Nights distribution is lopsided: Goa 1n / Jaipur 6n means Goa is effectively one evening (arrive Day 1 14:00, fly out Day 3 morning) - the 'Beach & Seafood' Day 2 framing implies a real Goa stay that the schedule doesn't deliver.
• Goa Day 2 zigzags: 'Saree Store in Margao' (15.272,73.959) at 09:30 is ~35km south of 'Burger Factory' (15.622,73.731) at 12:30 - long drive for a single shopping stop with no narrative reason.
• Prose is generic AI filler: 'Instagram-worthy moments' (Day 2), 'palaces, markets, and golden-hour desert light' (Day 4 - Jaipur is not desert), 'Desert Luxury Escape' (Day 5 - Jaipur is semi-arid, not desert).
• Day 3 has only one post-flight activity (City Palace 15:30, 90min) - a 4.5h flight + single short stop is thin even for slow pace.
• Flight cost asymmetry: Bengaluru->Goa ₹4,248 vs Jaipur->Bengaluru ₹13,554 (3.2x) with no fare-window justification visible - poor budget-aware sequencing.

Issues: Goa -> Jaipur mode is internally contradictory: logistics row says train 1742 min (~29h) at ₹2,575; Day 3 prose and activity card say flight 4.5h / 330min. The trip cannot be both.; If the realized leg is the 29h train, it is unreasonable for a 7-day slow-pace couple trip and burns Day 3 entirely; if it is the flight, the logistics row is wrong.; Day 5 routes through 'Boutique Luxury Retreat Rajasthan' at (26.568,72.346) - ~350km from the Nahargarh Fort bookends scheduled the same morning and evening - geographically impossible round-trip.; Goa 1-night / Jaipur 6-night split is lopsided for a 7-day 'heritage' couple trip; would expect 3-3 or 2-5 between two cities, or the originally requested 3-city Kerala loop.; Day 5 lists 'Boutique Luxury Retreat Rajasthan' at ₹300 for a 300min slot - implausible price extraction for a boutique luxury retreat day experience.; Activities total ₹8,650 across 7 days is light for mid-tier; several days show no cost line (Day 5 Nahargarh trek, Day 6 Jal Mahal + Amber village, Day 7 Hawa Mahal sunrise) - likely incomplete pricing rather than free entries.; ₹13,554 return flight from Jaipur dominates the flight subtotal (₹20,377); a re-routed direct Kerala flight from Bengaluru would have been a fraction of this cost.

ladakh-adventure-friends BROKEN heavy 1.8/5

India from Delhi · 9d · kind ["heritage"] · friends · budget premium · saves none

Catastrophic intent collapse: user requested Leh/Nubra/Pangong adventure, got Goa+Jaipur+Udaipur tourist circuit with zero Ladakh content and zero acclimatization safety.

Itinerary realism 2 · Saved-place utilization 1 · Route & transport 1 · Budget & cost 3  ||  Vibe/kind fit 1 · Companion fit 3 · Pace fit 2 · Prose quality 2 · Day balance 2

Good

• Hotels chosen are real, well-rated 4.6-4.7 properties (ITC Grand Goa Arossim, Jai Mahal Palace Jaipur, Taj Fateh Prakash Palace Udaipur) and consistent with the premium tier.
• Within Jaipur and Udaipur, the chosen monuments are real and canonical (Hawa Mahal, Amber Fort, City Palace Udaipur, Jagdish Temple, Nahargarh Fort, Lake Pichola).
• Goa->Jaipur and Udaipur->Delhi are correctly modeled as flights (sane mode given distance and absence of direct rail).

Bad

• INTENT COLLAPSE: user picked Leh/Nubra/Pangong (a high-altitude Ladakh adventure) and the actual trip is Mumbai (1n) -> Goa (3n) -> Jaipur (3n) -> Udaipur (2n). Zero Ladakh, zero adventure content, zero overlap with the requested geography. A traveler would reject this on sight.
• SAFETY FAILURE: a premium Ladakh adventure in late June requires 2+ acclimatization days in Leh before Nubra (Khardung La 5,359m) or Pangong (4,350m). The trip ignores this entirely because it never goes to Ladakh.
• MONSOON MISMATCH: Goa 3 nights and Udaipur 2 nights on 21-26 June lands squarely in SW monsoon onset — 'Anjuna Beach & Flea Market' (flea market closes in monsoon), 'Lake Pichola Sunset Boat Ride', 'Sajjangarh sunset photography', 'Mangrove Creek Boat Ride' are all weather-vulnerable choices for these dates.
• GEO-IMPOSSIBLE COORDINATE: 'Mangrove Creek Boat Ride' on Day 2 (Goa) is plotted at (12.250, 92.767) — that is Andaman Islands (Port Blair area), not Goa (which is ~15.5N, 73.8E). Either the wrong place entirely or a fabricated coordinate.
• DELHI->MUMBAI->GOA dogleg: routing the user via Mumbai for 1 night before Goa is absurd from Delhi when direct DEL->GOI flights exist. Adds a wasted flight (₹11,856), a wasted hotel night (₹18,497 at ITC Grand Central), and a 10-hour Konkan train Mumbai->Goa eating Day 2.
• TRIP NAME CONTRADICTION: 'North India Adventure: Goa to Jaipur' — Goa is not North India and there is no adventure content; nomenclature is internally incoherent.
• BREACHES 'avoidances: early_starts': Jagdish Temple at 06:30, Hawa Mahal at 07:30, Sunrise Photography Session at 06:00, Amber Fort at 07:30, Day 5 train at 07:00, Day 1 flight at 06:00, Day 9 flight at 08:00 — at least 7 pre-09:00 starts across 9 days despite explicit user avoidance.
• WRONG TRIP_KIND HONORED: user picked 'heritage' (not adventure per the fixture's destinations, but the analyst-intended adventure context); the produced trip is a generic sightseeing circuit with one 'Boutique Luxury Retreat Rajasthan' day at (26.568, 72.346) — that coordinate sits in the Thar near Jodhpur/Osian district, far from the Jaipur->Udaipur axis, with no transit segment showing how anyone gets there or back. Day 7's '600min' luxury-retreat block is filler.
• DUPLICATE STOP: 'Johari Bazaar & Bapu Bazaar' appears on both Day 6 (09:30, ₹3,000) and Day 8 (14:00, ₹2,000) — same Jaipur market, padded twice.
• Day 1 only has lunch-time arrival + flea market + dinner — no real first-day content; Day 9 is a single transit line. Effective trip is ~7 working days for a 9-day premium ask.
• Restaurant 'Antares by Goan Chef Sarah Todd' — Antares in Vagator is by Sarah Todd, but the phrasing 'Goan Chef Sarah Todd' is factually wrong (Sarah Todd is Australian); minor but a credibility tell.
• Pace was 'packed' yet most days have 2-3 anchors with large midday voids (e.g. Day 3: 09:30 flight lands ~14:00, then 15:30 + 17:15 = two short stops, nothing after 18:30). Doesn't read packed.

Issues: Delhi->Mumbai->Goa is a needless dogleg; Delhi->Goa direct is the obvious premium-tier choice and avoids both an extra hotel night and a 10h Konkan train.; Mumbai->Goa modeled as 600-min train (₹636) on a premium itinerary contradicts the budget tier; premium would fly or use Vande Bharat-class.; Goa->Udaipur Day 3 transit entry is listed under Day 3's day-card but build-overview segments show Goa->Jaipur (350min flight) — internal inconsistency between day plan and logistics segments.; Day 7 'Boutique Luxury Retreat Rajasthan' is plotted at (26.568, 72.346), ~200km west of Jaipur, with no transport segment to get there or back; trip silently teleports.; Zero Ladakh transport (which is the actual ask: Delhi->Leh flight, Leh->Nubra over Khardung La ~5h drive, Leh->Pangong ~5h drive via Chang La). None of this exists.; Total per-pax ₹2,58,428 is plausible for a premium 9-day domestic trip, but the spend is misallocated: ₹49,964 on flights for a route that should not include Mumbai dogleg, ₹30,000 in a single Day 7 'Boutique Luxury Retreat' activity block with no rationale.; ₹0 activity-cost entries on Day 3 (City Palace Lake View, Evening Lakeside Stroll) and Day 8 (Amber Fort Viewpoint Photography) and Day 5 (City Palace Complex listed at ₹500 vs Day 3's ₹0 'City Palace Lake View' — same complex priced inconsistently).; Sajjangarh sunset entry priced ₹3,000 vs actual entry ~₹100 INR — order-of-magnitude inflation.; Lake Pichola sunset boat ride ₹5,000/pax is at the very top of even premium-charter pricing; common shared sunset boat is ₹400-800. Plausible only if private charter, not stated.

montenegro-nature-couple BROKEN heavy 2.0/5

Montenegro from Delhi · 7d · kind ["nature"] · couple · budget mid · saves moderate

Geo-impossible Tivat->Sveti Stefan 'flight' + zero hotels + backwards Day 5 schedule sink an otherwise nature-fit Montenegro plan

Itinerary realism 3 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 4 · Companion fit 3 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Day 2 'Bay & Medieval Towns' geographically tight - Kotor Bay boat (42.425,18.769), Perast (42.486,18.699), Kotor Fortress (42.422,18.775) and Kotor Old Town all sit within Boka Bay, sensible single-day cluster
• Day 3 Lovcen + Cetinje + Njegos Mausoleum is a legit classic coastal-Montenegro nature/heritage daytrip, well-themed for a 'nature' couple trip
• Day 5 Durmitor National Park + Black Lake loop is a real and appropriate nature anchor for the trip kind
• Honors the early_starts avoidance well (Day 1 17:30, Day 2 09:00, Day 3 09:00) except Day 4 08:00 which is borderline and Day 7 07:30 farewell
• Uses 3 of the user's actual saves as named stops: Sveti Stefan (Day 1, Day 7 x2), Kotor Bay (Day 2 boat tour), Perast (Day 2)

Bad

• GEO-IMPOSSIBLE TRANSPORT: 'Tivat -> Sveti Stefan | mode=flight | 240 min | Rs 15,000' - these towns are ~25km apart on Montenegro's coast, ~40min by car/taxi. There is no flight, no airport at Sveti Stefan, and 240min for a 25km hop is absurd. This is a planning failure a normal user would immediately reject.
• STRUCTURAL FAIL: zero hotels persisted across a 7-night trip - hotels_total=0. User has no place to sleep and no accommodation cost, breaking the 'mid' budget tier promise and the trip plan entirely.
• DELHI->TIVAT 240min/4h flight is implausible - no direct flight exists; realistic Delhi->TIV via Istanbul/Vienna/Belgrade is 12-16h with a layover. Single 4h leg at Rs 15,000 misleads on both time and price (real fare ~Rs 45-70k one-way mid-season).
• Day 5 SCHEDULE IS BACKWARDS: 'Durmitor National Park at 09:00 510min' (ends 17:30), then 'Bobotov Kuk summit at 14:30 180min' (overlaps Durmitor by 3h), then 'Black Lake loop at 17:45 255min' (ends 22:00 - a 4.25h hike starting at dusk in mountains). Physically impossible ordering and a hike no normal couple does after sunset.
• INTRA-DAY OVERLAPS Day 3: 'Kotor Serpentine at 12:45 105min' overlaps 'Cetinje Old Town at 13:00 150min' - cannot be in both places at once, and they are ~30km apart.
• ROUTING ABSURDITY: Sveti Stefan -> Nevidio Canyon (Day 4) -> back to Sveti Stefan, then Sveti Stefan -> Durmitor (Day 5, ~250km north) -> back to Sveti Stefan, then Sveti Stefan -> Herceg Novi (Day 6, ~90km north) -> back. 6 nights pinned to Sveti Stefan turns this into massive daily drives - Durmitor alone is a 4-5h one-way drive. Should have been a Kotor base + a Zabljak overnight for Durmitor.
• Missed/ignored saves: Cetinje appears as 'Cetinje Old Town' (OK), but Blue Cave (saved, in Boka Bay near Herceg Novi) is included Day 6 but Portonovi is dropped onto wrong coords (42.790,19.251 is inland near Niksic, not Portonovi which is ~42.46,18.55). Bar, Petrovac, Ulcinj saves entirely ignored despite Petrovac being a 5min drive from the Sveti Stefan base.
• REQUESTED CITIES MISSING: user's intended cities were Kotor + Perast + Budva. Trip route is Tivat (1n) + Sveti Stefan (6n) - Budva and Kotor as overnight bases are dropped. Kotor and Perast appear as day-trip stops only.
• Coordinate errors: 'Evening walk, Podgorica Old Town' (42.442,19.262) and 'Morača River viewpoint' (42.363,19.058) tacked onto Day 4 after Nevidio Canyon (42.991,19.068) - Podgorica is a ~90min detour from any Sveti-Stefan-bound return route, with arrival at 18:00 implying late-night drive back to Sveti Stefan (~90km).
• Cost summary: Rs 50,335 total per person for 7n Montenegro with flights + activities is fictional - real all-in mid-tier should be Rs 1.4-2L per person. Activities total Rs 5,335 is also implausibly cheap for boat tours + canyon + 2 national parks.

Issues: Tivat -> Sveti Stefan tagged mode=flight (should be drive ~40min, ~25km) - geo-impossible; Delhi -> Tivat shown as 240min single leg at Rs 15,000 - no direct flight exists; real routing is via IST/VIE/BEG, 12-16h, Rs 45-70k mid-season; Sveti Stefan -> Delhi return shown as 240min flight, same issue as arrival - misleads on duration + cost; All three legs identically Rs 15,000 / 240min - templated, not real fare data; Base in Sveti Stefan for 6 nights forces ~250km Durmitor day-trip (impossible same-day with sane drive time) and ~90km Herceg Novi day-trip - should have split bases (Kotor or Budva + a Zabljak overnight for Durmitor); hotels_total=0 across 7 nights - structural fail, no accommodation booked or priced; Total Rs 50,335 per person for 7n international trip is impossibly low; real mid-tier Montenegro with international flights should be Rs 1.4-2L per person; Activities total Rs 5,335 omits paid items (Kotor Fortress entry is Rs 150 but Cetinje monasteries, Lovcen entry, Durmitor entry, Black Lake parking all under-counted; Nevidio Canyon Rs 3,000 looks like a single guide fee, but real canyoning packages run EUR 80-120pp = Rs 7-11k); Blue Cave boat tour priced at Rs 35 - implausibly low (real cost EUR 20-30 = Rs 1,800-2,700); No transparency on the budget tier: 'mid' should price ~3-star hotels at Rs 6-10k/night = Rs 42-70k accommodation alone, currently zero

Structural fails: hotels present; hotels priced (cost_summary.hotels_total > 0)

northeast-nature-friends BROKEN heavy 2.0/5

India from Kolkata · 8d · kind ["heritage"] · friends · budget mid · saves none

Catastrophic destination mismatch - asked for Guwahati/Shillong (NE India), got Mumbai/Goa/Jaipur/Udaipur with broken intra-route flights and a stray Margao saree-store stop.

Itinerary realism 3 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 2 · Prose quality 2 · Day balance 2

Good

• Within the (wrong) Udaipur day, the City Palace -> Saheliyon ki Bari -> rooftop sunset arc on Day 4 is coherent and uses real canonical Udaipur sights with plausible 09:30-17:15 ordering.
• Jaipur Day 6 anchors on real, well-known places (Hawa Mahal sunrise, City Palace, Jantar Mantar, Johari/Bapu Bazaar, Rawat Mishthan Bhandar for pyaaz kachori) in geographically tight Pink City coordinates (~26.92, 75.82).
• Kolkata -> origin-leg correctly modeled as a flight (175 min, ~Rs 11,100), which is the appropriate mode for the Kolkata <-> West/NW India hop.

Bad

• SEVERE PLANNING FAILURE: trip request explicitly named Guwahati + Shillong (NE India heritage with friends) and the produced itinerary contains zero NE India - route collapsed to Mumbai (1n) -> Goa (3n) -> Jaipur (2n) -> Udaipur (2n). No Kamakhya, no Umiam Lake, no Cherrapunji/Mawlynnong, no Shillong drive from Guwahati. The trip a normal user would reject on sight.
• Trip name 'India Tricolor: Goa to Rajasthan' actively confirms the misroute - it doesn't even pretend to be the NE heritage trip the user asked for.
• Day 3 'Flight to Lakes' fires a Goa -> Udaipur flight (225 min, Rs 4,500) but the build-overview transport spine has Goa -> Jaipur train (1742 min) then Jaipur -> Udaipur train (420 min) - itinerary-day transit contradicts the booked segments, so Day 3 arrives Udaipur with no corresponding selected flight.
• Day 5 is a single 23:00 overnight train Udaipur -> Jaipur (420 min) with literally zero other activity - that's not a 'slow' pace day, it's a dead day, and 23:00 -> 06:00 arrival violates the user's 'early_starts' avoidance for Day 6's 07:30 Hawa Mahal.
• Day 8 'Jaipur -> Kolkata' listed at 135 min in itinerary but selected flight in overview is 285 min Udaipur -> Kolkata at Rs 19,314 - segment city + duration + cost all disagree.
• Mumbai (1n) is a phantom city: Day 1 flies Kolkata -> Goa directly, Day 2 is in Goa, yet overview lists a Mumbai hotel (Royal Hometel Suites, Rs 5,670) and Kolkata->Mumbai + Mumbai->Goa train (600 min) segments - the user is billed for a Mumbai night they never spend.
• 'Saree Store in Margao' is pulled from the user's Panaji save and inserted as a 105-min Day 2 activity at coords (15.272, 73.959) - this is a single-shop stop, not a day-anchor; lazy padding.
• 'Boutique Luxury Retreat Rajasthan' on Day 7 at Rs 30,000 (coords 26.568, 72.346 - Jodhpur region, ~300km from Jaipur) is scheduled 13:00-15:00 between Amber Fort and Nahargarh Fort - geographically impossible as a 2-hour midday side trip from Jaipur.
• Activities total Rs 63,088 per person is inflated by that single Rs 30,000 'Accommodation' activity line - a hotel masquerading as an activity, double-counting against the Rs 19,048 hotels_total.
• Day 1 'Anjuna Beach & Flea Market' Rs 5, Day 2 'Antares' Rs 25, 'Brittos' Rs 8 - sub-Rs 50 entries on flagship Goa stops are meaningless placeholder costs, not budget-tier realism.
• Prose is generic AI filler: 'beach vibes', 'Instagram moments', 'Coastal Rhythms', 'Train Through Desert' - says nothing specific about the actual stops; 'Lunch Stop' literally named 'Lunch Stop' is the floor of effort.

Issues: Phantom Mumbai stopover: overview has Kolkata->Mumbai flight + Mumbai->Goa train + 1-night Mumbai hotel, but itinerary Day 1 flies Kolkata->Goa direct at 06:00 (violates early_starts avoidance too).; Day 3 itinerary flight Goa->Udaipur (225 min, Rs 4,500) is not in the selected_flights spine, which has Goa->Jaipur train (1742 min) instead - itinerary and bookings describe two different trips.; Day 5 overnight train 23:00 Udaipur->Jaipur arrives ~06:00 forcing an early-start Day 6 (Hawa Mahal 07:30) - violates user's explicit 'avoidances: early_starts'.; Day 8 itinerary lists Jaipur->Kolkata at 135 min, Rs 8,000; overview selected_flight is Udaipur->Kolkata at 285 min, Rs 19,314 - cost summary uses one, itinerary shows the other.; Most critically: ZERO Kolkata->Guwahati flight (the obvious and correct origin->NE gateway) and ZERO Guwahati->Shillong 3-4h drive - the entire NE India gateway logic was bypassed.; 'Boutique Luxury Retreat Rajasthan' Rs 30,000 logged as an activity line item (cat=Accommodation) on Day 7 - hotel cost double-counted into activities_total (Rs 63,088) on top of the Rs 19,048 hotels_total.; Rs 5 Anjuna Beach, Rs 25 Antares, Rs 8 Brittos - placeholder near-zero costs on flagship Goa stops; Antares (celebrity-chef restaurant) realistically Rs 2,500-4,000 pp.; Cost_summary total Rs 1,16,259 pp is incoherent with the displayed nights (Mumbai 1 + Goa 3 + Jaipur 2 + Udaipur 2 = 8) given the inflated Rs 30,000 Accommodation activity line.; Mumbai hotel Rs 5,670 charged for a night Day-1 itinerary skips entirely - real cash for a phantom stay.

rajasthan-heritage-family BROKEN heavy 1.8/5

India from Delhi · 8d · kind ["heritage"] · family_kids · budget premium · saves thin

Rajasthan heritage request collapsed into a Mumbai-Jaipur-Udaipur-Goa hodgepodge with absurd transport, broken geocodes, and a 'Hawa Mahal at Sunrise' scheduled for 13:00.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 2 · Companion fit 2 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Jaipur Day 2 anchor sequence (Amber Fort 07:30 -> Jantar Mantar 10:45 -> Albert Hall Museum 13:30) is geographically coherent and hits the canonical pink-city heritage trio.
• Udaipur Day 4 has a sane heritage shape with Jagdish Temple, Saheliyon-ki-Bari, Bagore-ki-Haveli, and a Lake Pichola sunset boat - all real, kid-friendly heritage stops.
• Hotel picks are genuine heritage/luxury properties (Jai Mahal Palace Jaipur, Taj Fateh Prakash Palace Udaipur) appropriate to the premium tier and family-with-kids profile.
• Jaipur -> Udaipur leg correctly routed as a train (420 min, ~Rs.498), not a flight - one of the few sane intra-Rajasthan choices.

Bad

• MAJOR PLANNING FAILURE: user asked for Rajasthan heritage circuit (Jaipur + Udaipur + Jodhpur). Flow dropped Jodhpur entirely and substituted GOA - beach state, not heritage, not Rajasthan. Days 5-8 (half the trip) are in the wrong region.
• Inserted a phantom Mumbai layover (1 night, ITC Grand Central, Rs.18,497/night) on the way from Delhi to Jaipur - Delhi->Jaipur is a 4hr drive or 1hr direct flight; no human routes Delhi->Mumbai->Jaipur. This is geographically absurd and burns Rs.13,428 on a useless detour flight.
• Day 1 contradiction: prose says 'Early drive to Jaipur (5-6h)' but the transit row is 'Delhi -> Jaipur 180min' AND the user's avoidances explicitly include 'early_starts'. Day 1 then schedules 'Hawa Mahal at Sunrise' at 13:00 - the name is internally contradictory (sunrise at 1pm).
• Day 2 'Boutique Luxury Retreat Rajasthan' at 16:00 lists coords (26.568, 72.346) which is near Jodhpur, ~330km from Jaipur - cannot be a 4-hour afternoon activity from a Jaipur base. The user's one real saved place is geo-mislocated and shoehorned as a Rs.30,000 'activity'.
• Day 6 'Mangrove Creek Boat Ride' lists coords (12.250, 92.767) - that is in the ANDAMAN ISLANDS, not Goa. Day in Goa contains a stop ~2000km away. Hallucinated/wrong-geocoded place.
• Day 2 violates 'early_starts' avoidance (Amber Fort 07:30) and Day 4 doubles down (Jagdish Temple 07:00) - core user preference ignored twice.
• Day 7 schedules 'Diaz Nightlife' at 20:30 for 120min on a family-with-kids trip - adult nightlife is not companion-appropriate. 'Lazy Dog Lounge Walk' at 14:00 also reads as adult-bar coded.
• Day 8 has only one activity (Calangute Beach Morning 07:30) - again violates early_starts, and 90 minutes is the entire 'departure day' with no actual transport row for Goa->Delhi (flight only exists in the logistics summary, not the itinerary).
• Udaipur -> Goa flight Rs.34,009/pax for a 255-min hop is high even at premium tier; cumulative flight spend Rs.81,481/pax is bloated by the unnecessary Delhi->Mumbai and Udaipur->Goa legs.
• Day 5 'Antares by Goan Chef Sarah Todd' is the only dinner anchor for the whole Goa block; Days 6/7 have no real lunch anchor and Day 6 has no dinner anchor - meal cadence is broken for a family trip.

Issues: Delhi -> Mumbai -> Jaipur is a phantom detour; Delhi to Jaipur is direct (1h flight or 4h drive). The Mumbai 1-night stop is unjustified for a Rajasthan heritage trip.; Udaipur -> Goa as a 255min flight for a Rajasthan-themed trip is non-sequitur; the originally-intended Jodhpur leg (Udaipur->Jodhpur is a 250km drive) was dropped.; Goa -> Delhi return flight Rs.18,120 appears in the logistics summary but has no corresponding transit row on Day 8.; Day 6 'Mangrove Creek Boat Ride' coords (12.250, 92.767) are in the Andaman Islands - impossible to do from a Goa base.; Rs.30,000 single-activity charge on Day 2 for 'Boutique Luxury Retreat Rajasthan' inflates the activities subtotal; it's a hotel/retreat, not a 240-min activity.; Rs.81,481/pax in flights driven largely by two unnecessary legs (Delhi->Mumbai Rs.15,426 and Udaipur->Goa Rs.34,009) - both avoidable with the intended Jaipur->Udaipur->Jodhpur->Delhi routing.; Total Rs.2,54,107/pax for an 8-day domestic premium trip is on the high side because of the bloated flight/hotel count (4 hotels in 8 nights including a useless Mumbai night at Rs.18,497).

switzerland-family-premium BROKEN heavy 2.3/5

Switzerland from Mumbai · 9d · kind ["alps"] · family_kids · budget premium · saves none

Real, kid-friendly Swiss alpine picks ruined by catastrophically inflated activity costs (Titlis ₹99k, GoldenPass ₹69k), a day plan whose geography contradicts its own booked route, and a Day-9 departure from the wrong city for the booked return flight.

Itinerary realism 2 · Saved-place utilization 4 · Route & transport 2 · Budget & cost 1  ||  Vibe/kind fit 4 · Companion fit 4 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Every named place is a real, recognizable Swiss landmark and genuinely kid-appropriate: Mt Pilatus cogwheel/cable car, Lucerne Chapel Bridge, Jungfraujoch via Jungfrau Railway, Titlis rotating cable car + ice caves, Lauterbrunnen Valley, Gruyeres cheese experience, Jet d'Eau, Lake Thun boat cruise - no hallucinated junk.
• The user's single saved place, Oberhofen Castle (Bern), is actually used on Day 8 as the lakeside-tour anchor - rare for a pure-fallback trip to hit the one save.
• Hotels are all real top-tier Swiss properties (Waldhotel by Burgenstock ₹92k, The Dolder Grand in Hottingen ₹76k, Beau-Rivage Geneve at Mont-Blanc ₹1.13L) with sane premium-tier per-night figures for a very-expensive market.
• Trip_kind 'alps' is well served end-to-end (Pilatus, Titlis Glacier, Jungfraujoch 'Top of Europe', Grindelwald First cliff walk); prose is punchy and concrete ('cogwheel up Mt Pilatus, rotating cable car down', 'Photography gold').

Bad

• Activity costs are 10x+ inflated and would mislead the traveler badly: 'Titlis Rotating Cable Car' ₹99,000 (real ~CHF 96 / ~₹9.5k), 'Grindelwald First Cliff Walk' ₹54,450 (real ~CHF 72 / ~₹7k), 'Lucerne -> Geneva via GoldenPass' ₹69,300 and 'Train: Gruyeres -> Grindelwald' ₹34,650 (scenic/regional fares are a fraction), 'Lake Thun Boat Cruise' ₹44,550, 'Oberhofen Castle' entry ₹27,500 (real CHF 12 / ~₹1.2k). The ₹3,60,171 per-person activity line is not a real Swiss cost - smells like a paise/rupee unit bug.
• ROUTE INCOHERENCE: build-overview says Lucerne(3n) -> Zurich(3n) -> Geneva(3n) but the day plan runs Lucerne(D1-2) -> Geneva(D3-4) -> Zurich(D5-6) -> Lucerne-area(D7-9). The stated transport segments 'Lucerne -> Zurich | drive 48min' and 'Zurich -> Geneva | train 168min' are never executed in the day-by-day.
• Day 9 'Lucerne -> Zurich Airport' departs from Lucerne, but the booked return flight is Geneva->Mumbai per the logistics - the traveler is in the wrong city to make their flight. A normal user would miss the plane.
• Day 4 prose calls Gruyeres a trip into the 'French Alps' - Gruyeres is in Switzerland (canton Fribourg), not France. Factual error.
• Day 4 also claims 'Train: Gruyeres -> Grindelwald | 135min' - the real journey is 3+hrs with multiple changes via Bern/Interlaken Ost; and the implied return to Geneva that night never gets a transport leg.
• Avoidance 'early_starts' is violated: Day 6 'Zurich -> Lauterbrunnen Train' is scheduled at 07:30, and the Jungfraujoch day trip from a Zurich base (~2.5h each way) is a punishing day for a family with kids.
• Mumbai -> Lucerne is labeled 'flight' in logistics (Lucerne has no airport); Day 1 narrative correctly handles it as fly-to-Zurich-then-train, but the segment label is wrong.
• Day 3 'Jet d'Eau & Lake Geneva' coords (46.412, 6.160) sit ~23km south of the real Jet d'Eau at (46.207, 6.156) - near the French border, not in Geneva.
• Day 9 is degenerate: one 75-min airport transfer (₹198) for a premium 9-day family trip closer - no meal, no buffer, no breakfast.

Issues: Route/day-plan mismatch: build-overview says Lucerne -> Zurich -> Geneva, but days run Lucerne -> Geneva -> Zurich -> Lucerne-area; the booked 'Lucerne -> Zurich | drive 48min' and 'Zurich -> Geneva | train 168min' segments never happen.; Day 9 departs 'Lucerne -> Zurich Airport' though the booked return flight is Geneva -> Mumbai - user is in the wrong city to make their flight.; Lucerne -> Zurich logistics segment is mode=drive 48min ₹80 in the world's best rail country with sub-1hr direct SBB trains every 30min - should default to train.; Day 4 'Train: Gruyeres -> Grindelwald 135min' understates the real 3+hr multi-change journey via Bern/Interlaken Ost.; Day 7 (Titlis, Engelberg near Lucerne) and Day 8 (Oberhofen/Thun, Bernese Oberland) are run from a Zurich base per the route, both 1.5-2.5h+ each way, with no transit legs shown.; Mumbai -> Lucerne mislabeled as direct 'flight' - Lucerne has no airport.; Titlis Rotating Cable Car ₹99,000 - ~10x the real ~CHF 96 (~₹9.5k).; Lucerne -> Geneva via GoldenPass ₹69,300 - a 2nd-class scenic-rail ticket is ~CHF 80-120 (~₹8-12k).; Grindelwald First Cliff Walk ₹54,450 - real first-cliff-walk+cable-car ~CHF 72 (~₹7k).; Lake Thun Boat Cruise ₹44,550 and Train Gruyeres -> Grindelwald ₹34,650 - both an order of magnitude too high.; Oberhofen Castle entry ₹27,500 (real CHF 12, ~₹1.2k) and Gruyeres Cheese Experience ₹14,850 (real CHF 7, ~₹700) - castle/cheese-house admissions are off by ~20x.; Montreux Lakeside ₹11,880 charged for a 90-min free lakeside stroll - should be ₹0.; Mumbai-Zurich return flights ₹15k each way are suspiciously CHEAP for premium-tier April (realistic ₹40-60k one-way), pointing to inconsistent pricing layers.; Resulting ₹3,60,171 per-person activity total dominates the ₹12.35L grand total and is not a realistic Swiss-trip activity spend even at premium tier - real number is ₹40-70k including all alpine railways.

tamilnadu-heritage-parents BROKEN heavy 1.5/5

India from Chennai · 6d · kind ["heritage"] · with_parents · budget mid · saves none

Tamil Nadu heritage circuit for parents collapsed into a Goa beach + Jaipur trip with early starts and a hike, ignoring every part of the request.

Itinerary realism 2 · Saved-place utilization 1 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 1 · Companion fit 1 · Pace fit 2 · Prose quality 2 · Day balance 2

Good

• Jaipur Pink City day (Day 4) is internally coherent: City Palace 09:00 -> Jantar Mantar 11:30 -> Bapu Bazaar 14:30 -> Hawa Mahal golden hour 17:00 are all real, close together (~26.92, 75.82), and reasonably paced for parents.
• Goa -> Jaipur leg correctly modeled as a flight (1742min train was originally listed but the day-3 itinerary uses a 330min flight at 09:00, which is the geo-sane choice).
• Hotels are real, sub-3-star-budget-friendly properties (Bloom Hotel Dona Paula 1842/night, Hotel Grand Lotus Inn 1293/night) which is at least consistent with a mid tier on lodging.

Bad

• SEVERE PLANNING FAILURE: user asked for a Chennai -> Pondicherry -> Madurai heritage temple circuit; the flow shipped Delhi (1n) -> Goa (3n) -> Jaipur (2n). Zero Tamil Nadu content. No Meenakshi Amman, no Mahabalipuram, no Pondicherry French Quarter, no Brihadeeswarar. The trip is a different trip.
• Day 1 'Anjuna Beach & Flea Market' at 17:00 + 'Britto's Beach Shack' at 19:45 is a beach-party itinerary - a direct miss for a with_parents heritage request and the opposite of 'gentle/accessible'.
• Day 2 'Mangrove Creek Boat Ride' coords (12.250, 92.767) are in the Andaman Islands, not Goa - hallucinated/geo-broken stop; impossible to do on a Goa day without a flight.
• Day 2 07:30 boat ride and Day 5 'Amber Fort Early Start' 07:30 directly violate the explicit avoidance ['early_starts'].
• Day 5 'Nahargarh Fort Sunset HIKE' at 16:30 (120min) is an extreme/strenuous activity flagged in the title as a hike - inappropriate for with_parents companion.
• Day 5 inserts 'Lal Tibba' at coords (30.477, 78.041) - that is in MUSSOORIE, Uttarakhand, ~900km from Jaipur. Teleport stop in the middle of a Jaipur day; clearly pulled in from the user's Mussoorie saves and dropped without geography awareness.
• Day 3 lists 'Check into Boutique Luxury Retreat' as a 30,000 INR 'activity' (150min) - check-in is not an activity, and 30k as a single line item is nonsense (also coords 26.568,72.346 are near Jodhpur, not Jaipur).
• Route mode for Goa -> Jaipur in the overview claims a 1742-minute (29-hour) TRAIN at 2,575 INR - that's borderline acceptable distance-wise but contradicts the Day 3 itinerary which shows a 330min flight; one of the two is wrong.
• Initial Chennai -> Delhi -> Goa routing forces two flights (655min, 25,360 INR) to reach a destination ~1500km direct from Chennai, when a direct Chennai -> Goa flight (~90min) exists. Gateway logic is broken.
• Trip name 'Goa & Jaipur in 6 Days' acknowledges the wrong cities; itinerary subtitle 'Coastal Flavors' and 'Sunset photography' are tonally wrong for a heritage-with-parents brief.
• Activities total 43,600 INR for 6 days at mid tier with parents is high-skewed by the 30k 'check-in' line; remove that and activities are ~13.6k which is fine, but the cost summary still reports 43,600 - the per-person total of 99,017 is therefore inflated/misleading.

Issues: Chennai -> Delhi -> Goa double-flight gateway when direct Chennai -> Goa exists (~90min, ~5-7k INR); Goa -> Jaipur listed as 1742min train in overview but 330min flight in Day 3 itinerary - contradiction; No airport-transfer or arrival-day buffer on Day 1; lands and goes straight to Anjuna 17:00; Day 2 'Mangrove Creek Boat Ride' coords (12.250,92.767) are Andaman Islands, not Goa - geo-impossible inclusion; Day 5 'Lal Tibba' coords (30.477,78.041) are Mussoorie, ~900km from Jaipur - geo-impossible inclusion; Day 3 'Check into Boutique Luxury Retreat' priced 30,000 INR as a 150min 'activity' - check-in is not an activity and the figure dominates the activities total; Activities total 43,600 INR is misleading because of the 30k check-in line; real activity cost is ~13.6k; Per-person total 99,017 INR for 6 days domestic mid-tier with parents is acceptable as a number but the composition is wrong (47,305 flights for a should-be-short-haul TN trip); Hotels total 8,112 INR across 5 nights (2 hotels at ~1.3-1.8k/night) - low end of mid tier, not luxury but workable; 2-star territory could disappoint parents; Goa -> Jaipur train at 2,575 INR for 29h is unrealistically cheap for a sleeper berth and the overview-vs-day-3 mode contradiction makes the line untrustworthy

thailand-beach-budget BROKEN heavy 2.0/5

Thailand from Delhi · 7d · kind ["beach"] · friends · budget budget · saves rich

Drops Koh Samui entirely, barely uses Krabi's 45 saves, and ships zero hotels on a budget trip

Itinerary realism 3 · Saved-place utilization 2 · Route & transport 2 · Budget & cost 1  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 2 · Prose quality 3 · Day balance 2

Good

• Bangkok side actually uses real saved places: Wat Arun, Grand Palace, IconSiam, Chao Phraya Cruise Dinner, CHATUCHAK MARKET, Khao San Road, Tichuca Rooftop Bar are all from the user's saves
• Transport modes are geographically sane — Delhi->Krabi and Krabi->Bangkok are both flights (no drives across the Andaman/Gulf)
• Day 4 cleanly anchors the city handoff: Krabi->Bangkok flight at 09:00, then temple-cluster (Grand Palace, Wat Arun) into rooftop sunset (Tichuca) — coherent arrival day
• Damnoen Saduak appears exactly once on Day 6 despite being saved under both Bangkok and Ratchaburi — no duplicate

Bad

• Koh Samui was in the requested cities AND the user has 7 Koh Samui saves (Fisherman Village Night Market, Crystal Bay, Elephant Sanctuary, Pig Island, Lamai Night Market, Overlap Stone, Six Senses) — none of these surface; the third island leg is silently dropped
• Krabi is the user's #1 saved city globally (45 saves) and only 4 saves are used across 4 nights: Aonang Beach, Khao Ngon Nak Viewpoint, Klong Thom Hot Springs, Aonang Walking Street. Missed: 89 Cafe, Aonang Center Point, Aonang Landmark Night Market, Blue Dragon, Boogie Bar, Cabaret Show, Cafe 8 98, Catalunya, Chalita Restaurant, Chicken Island Tour, Chock Dee Handmade Jewellery, Emerald Blue Pool, Indian Corner, Khaotong Hill Viewpoint, Kodam Kitchen, plus ~25 more. James Bond Island Tour, Railay Beach, Tiger Cave Temple are AI inventions when there are 41 user-saved Krabi places sitting unused
• Phi Phi gets zero coverage despite 6 saves (Maya Bay, Pileh Lagoon, Monkey Beach, Tom Sai Beach, Koh Dalum, Phi Phi Islands) — these are reachable as a day-tour out of Krabi/Aonang and were the obvious beach-tier-utilization move
• Hotels are completely missing (hotels_total=0, cost_summary.hotels=NULL) — for a 7-night trip on budget tier this is a fundamental persistence failure, not a rounding error; the cost summary lies about total cost
• User explicitly avoided early_starts but Day 2 opens at 07:30 (James Bond Island) and Day 6 opens at 07:30 (Damnoen Saduak) — the avoidance was ignored on two of seven days
• Day 6 stacks 5 activities (Damnoen Saduak, Chinatown Food Market, Pad Thai Thip Samai, Chao Phraya Cruise Dinner, Khao San Road) ending at 23:00 — that is packed, not balanced, and conflicts with the stated pace=balanced
• Delhi->Krabi flight is ₹33,696 (one-way) and Bangkok->Delhi is ₹33,477 — total ₹76,676 in flights alone is rich-tier pricing, not budget-tier (a realistic budget DEL-BKK round-trip in March is ₹25-35k inclusive)
• Day 7 has a single activity (Chatuchak 09:00-11:00) and ends at airport transit 12:00 — under 3 hours of content on the last day reads like padding rather than a proper morning
• Trip kind = beach with only friends but 4 of 7 days are temples/markets/shopping — beach content is essentially Day 1 (Aonang) + Day 2 (Railay); the trip drifts into a Bangkok culture week

Issues: Koh Samui was in the requested cities list but does not appear in route, transport segments, or itinerary days — silently dropped without a Krabi->Koh Samui ferry/flight or any acknowledgment; Bangkok->Delhi return flight at ₹33,477 is the same price as the outbound — no obvious issue, but combined with the missing third leg (Koh Samui) the routing is incomplete relative to intent; hotels_total = 0 and hotels array empty — 7 nights of accommodation are unbudgeted; total cost ₹1,05,760 understates real trip cost by an estimated ₹15-35k (budget hotels in Krabi + Bangkok); Flights ₹76,676 per person for DEL->KBV->BKK->DEL is high for budget tier — March DEL-BKK round-trips typically clear ₹25-40k on IndiGo/Thai Lion; the auto-selected fares look like flexible/premium-economy not the cheapest in-budget option; Activities ₹29,084 over 7 days is fine, but cost_summary cannot be trusted when hotels are missing — user sees a misleading 'total' that is missing the largest variable cost line for a budget trip

Structural fails: hotels present; hotels priced (cost_summary.hotels_total > 0)

uttarakhand-hills-friends BROKEN heavy 1.8/5

India from Delhi · 6d · kind ["hill-stations"] · friends · budget mid · saves rich

Hill-station trip hallucinates a Mussoorie airport and bolts on a Goa beach leg the user never asked for, ignoring 14 of 19 Uttarakhand saves.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 1 · Budget & cost 2  ||  Vibe/kind fit 1 · Companion fit 3 · Pace fit 2 · Prose quality 2 · Day balance 2

Good

• Day 2 is the one bright spot: Kempty Falls, Jharipani Falls, and Cloud's End are all real, all from the user's Mussoorie saves, and form a coherent waterfall/forest theme.
• Day 1 correctly uses Camel's Back Road and Landour Bakehouse from the user's saves and pairs them with a sensible arrival-day pace.
• Lal Tibba on Day 3 morning is a real saved viewpoint and an appropriate light-touch activity before a transit day.
• Delhi -> Mussoorie segment correctly uses bus (222 min, Rs 221) instead of an impossible flight, respecting the 'no airport in the hills' reality on the inbound leg.

Bad

• INTENDED-VS-ACTUAL FAILURE: user asked for Mussoorie + Rishikesh (hill stations); flow shipped Mussoorie + GOA (beach). Rishikesh - the second intended city with 5 dedicated saves (Char Dukan, Jaberkhet Reserve, Kanatal Forest, Landour Bakehouse, Sister Bazaar) - is completely dropped. This is a categorical planning failure, not a near-miss.
• GEOGRAPHICALLY IMPOSSIBLE TRANSPORT: 'Mussoorie -> Goa | mode=flight | 325 min | Rs 16,070' and Day 3 'Mussoorie -> Goa flight at 09:00 / 300min'. Mussoorie has no airport. Nearest is Jolly Grant (DED) ~60km away via 1.5-2h road. The leg silently assumes a non-existent Mussoorie airport.
• Trip violates the 'hill-stations' trip_kind: half the nights (2 of 5 city-nights) are on a beach 2,000km away. A 6-day hill-stations trip with a forced beach pivot is a vibe failure no normal traveler asked for.
• Day 5 'Lazy Dog Lounge' is geo-tagged at (32.252, 77.187) which is Himachal Pradesh (Manali region), not Goa. Either a wrong place was picked or the coordinates are corrupted - either way the day's geo is broken.
• 'Saree Store in Margao' (15.272, 73.959) is in South Goa but slotted into Day 4 between Baga (North Goa) and Antares (Vagator/North Goa) - a ~40km southbound detour mid-afternoon then back north for dinner. Real driving plan would be 2.5h+ wasted on the road.
• 9 of 14 Mussoorie saves unused (Bhadraj, Bhatta Falls, Budha Temple, Dalai Hills, Dehradun, George Everest's House, Gun Hill, Kimadi Falls, Surkhanda) and 4 of 5 Rishikesh saves unused (Char Dukan, Jaberkhet Reserve, Kanatal Forest, Sister Bazaar). On the user's RICHEST domestic destination, the itinerary used only 6 of 19 relevant saves.
• Day 1 violates 'early_starts' avoidance: a 06:00 bus departure with 270min transit lands the user at Camel's Back by 11:15 - the entire morning was an early start the user explicitly said no to.
• Day 2 also violates 'early_starts': Kempty Falls at 07:30 after a Day 1 that ended late.
• Costs are inconsistent: Landour Bakehouse listed at cost=Rs 4 (typo/parse failure for Rs 400?), Mussoorie->Goa flight in itinerary shows cost=Rs 6,000 but logistics segment shows Rs 16,070 - two different prices for the same leg.
• Hotels: only Goa hotel is listed (Bloom Hotel Dona Paula). No Mussoorie hotel for the 3 nights there - the cost summary's Rs 3,684 hotels total covers 2 Goa nights only, meaning 3 Mussoorie nights are unpriced/unbooked.
• Prose is generic AI filler: 'settle into mountain town, evening exploration', 'Coastal exploration, beachside dining, local shopping' - templated, no editorial voice, no specifics tied to chosen places.
• 'Antares by Goan Chef Sarah Todd' - the restaurant is actually run by Sarah Todd (Australian chef) and is not Goan-chef-owned; the descriptor is wrong.

Issues: Mussoorie -> Goa modeled as direct flight at 325 min for Rs 16,070 - Mussoorie has no airport; real path is Mussoorie -> Dehradun (DED) road transfer ~2h then DED -> GOI flight ~3h with a layover (no direct DED-GOI exists year-round); itinerary collapses this to a single fictitious leg.; Day 3 in-itinerary 'Mussoorie -> Goa flight' at 09:00 with 300min duration conflicts with the logistics block's 325min - inconsistent durations for the same segment.; Goa -> Delhi flight 155 min / Rs 12,240 is plausible in isolation but the entire Goa leg shouldn't exist.; No transport segment for the Mussoorie -> Dehradun airport transfer required to make even a fictional onward flight work.; Day 5 Lazy Dog Lounge geo at (32.252, 77.187) is Himachal - either wrong place or impossible 2,500km mid-day teleport from Palolem (15.359) to Manali region.; Mussoorie -> Goa flight price discrepancy: Rs 16,070 in logistics segment vs Rs 6,000 in Day 3 itinerary activity row - two sources of truth disagree.; Landour Bakehouse cost=Rs 4 - almost certainly a parse/decimal error (real spend Rs 400-800/head); displayed as Rs 4 is misleading.; Hotels total Rs 3,684 = 2 nights at Bloom Hotel only; the 3 Mussoorie nights have no hotel selected or priced, leaving a real-world ~Rs 6,000-12,000 gap unaccounted for at mid tier.; Per-pax total Rs 43,379 dominated by Rs 28,531 in flights for what should be a Delhi-Mussoorie road trip - the bogus Goa pivot accounts for ~Rs 28,310 of the flight spend the user didn't need or ask for.; Mid tier flight Rs 28,531/pax for a 6-day domestic hill trip is wildly over-budget vs the real trip (Delhi-Dehradun bus + Mussoorie cabs would land under Rs 3,000 of transport).

Structural fails: no car/drive on origin (intl) leg

vietnam-mix-friends BROKEN heavy 2.3/5

Vietnam from Delhi · 8d · kind ["mix"] · friends · budget mid · saves thin

Phu Quoc (9 of 10 saves) entirely omitted; Ha Long Bay Cruise duplicated across Day 2 evening + full Day 3; Day 4 train Hanoi->HCMC mislabeled as 7.4h when it is a ~30h+ haul.

Itinerary realism 2 · Saved-place utilization 2 · Route & transport 2 · Budget & cost 3  ||  Vibe/kind fit 2 · Companion fit 3 · Pace fit 3 · Prose quality 3 · Day balance 2

Good

• Hanoi -> Ho Chi Minh City correctly modeled as a flight (130 min, ~₹18k) in the transport segments — appropriate for the ~1150km N-S hop.
• Day 5 surfaces the user's one HCMC save (Vườn Thiên Giới) alongside Cu Chi Tunnels — at least one named saved place is honored.
• Day 7 day-trip pairing (Golden Bridge / Ba Na Hills morning, Marble Mountains afternoon) is geographically coherent out of Hoi An.

Bad

• Phu Quoc utilization = 0/9: the user's densest saves cluster (Grand World, Hon Thom Cable Car, Kiss of the Sea, Sunset Town, VinWonders, Starfish Beach, Snorkeling Tour, Sunset Bar, Oc Sen) is the clearest signal in the trip and the itinerary never visits Phu Quoc at all — replaced with Hoi An, which has zero saves.
• Day 3 'Ha Long Bay Cruise' (570min, ₹14,400) is a duplicate of the Day 2 Ha Long Bay Cruise entry (already 330min, ₹14,400 at 15:30) — same coords (20.910, 107.184), same name, double-charged. Either a 2-day overnight cruise should be ONE item or it's a hallucinated repeat.
• Day 4 prose says 'Train from Hanoi to Ho Chi Minh City (7.4h overnight or day)' — the Reunification Express Hanoi-HCMC is ~30-36 hours, not 7.4h. The transit row also says 450min (7.5h) which is geographically impossible (~1700km rail).
• Day 6 prose says 'Train from Ho Chi Minh City to Hoi An (7.9h overnight or day)' — HCMC->Da Nang train is 16-19h, not 7.9h/475min. Both long-haul south-north train legs are time-collapsed to ~half the real duration.
• Hanoi gets 1 night allocated but the itinerary spends Days 1-4 there (Days 2-3 are Hanoi/Ha Long Bay activities) — nights-vs-activity-days are out of sync; the user would arrive in HCMC with no hotel for nights they'd already booked.
• Day 1 'Delhi -> Hanoi (Flight) at 06:00' violates the user's stated `early_starts` avoidance — a 6am intl departure means a 3-4am airport arrival.
• Day 7's 'Golden Bridge — Ba Na Hills at 07:30' also breaks early_starts; Day 6 06:00 train departure is the worst violation.
• Hoi An hotel 'Wild Bear Inn' is in neighborhood 'Conner Heights' — likely a hallucinated/typo'd neighborhood name (should be 'Cẩm Châu' or similar); rating 3.9 is the lowest of the three and most expensive (₹6,798) — inverted for mid-tier value.
• Day 2 stacks Hanoi Train Street + Old Quarter + Loading T Café + Ha Long Bay Cruise departure at 15:30 — Ha Long Bay is ~2.5h drive from Hanoi, so a 330-min cruise starting 15:30 there is impossible without transit time baked in (no transfer leg shown).
• City selection ignores the saves entirely: with 9/10 saves in Phu Quoc the auto-picker should have surfaced Phu Quoc as a primary city — choosing Hoi An (0 saves) over Phu Quoc (9 saves) is a planning failure.

Issues: HCMC -> Hoi An labeled as 'train' 444min/₹888 — Reunification Express HCMC->Da Nang is 16-19h actual (~960-1140min); the duration is roughly half of reality.; Day 4 narrative claims 'Train from Hanoi to Ho Chi Minh City (7.4h)' — actual Hanoi->HCMC by rail is ~30-36h; this is a fundamental misrepresentation that would strand the user.; Phu Quoc is an island and was not visited at all — but if it had been, the trip would have correctly needed a flight (not drive) from HCMC; the omission is the planning failure, not the mode.; Day 2 'Ha Long Bay Cruise' at 15:30 starting from Hanoi (3h coach transfer typical) with no transfer leg modeled — start time is impossible.; Hanoi (1 night) but Days 2 and 3 are Hanoi/Ha Long Bay — nights allocation and activity-days are inconsistent (cruise itself is usually 1-night overnight on the boat, which isn't accounted for in the hotel stack).; Delhi->Hanoi at ₹48,456 one-way is high for mid-tier (typical ₹25-35k one-way DEL-HAN; ₹40-55k more typical for round-trip in shoulder season) — possibly inflated.; Day 3 Ha Long Bay Cruise ₹14,400 appears to be a duplicate charge of the Day 2 Ha Long Bay Cruise ₹14,400 — total ₹28,800 for a single 2-day cruise product is plausible only if explicit, but the line items make it look double-counted.; Hoi An hotel ₹6,798/night with 3.9 rating is worse-value than the Hanoi 4.8-rated ₹3,760/night and HCMC 4.8-rated ₹1,742/night — ranking inverted for mid budget.; Cu Chi Tunnels ₹1 — placeholder/sentinel cost rather than a real estimate (actual ~₹1,500-2,500 with transfer).; Activities total ₹40,244 includes the duplicated Ha Long cruise charge — real-world activities total is closer to ₹25-26k.