Cycling through all six. Tap any point to stop.

Measured on six things

AI travel concierge that books places via text or LINE app

Security

Security-header probe scores 2/12: no enforced HTTPS, no HSTS, no CSP, no frame protection, no referrer or permissions policy. SPF and DMARC are present. No SOC 2, ISO, or bug-bounty signals found.

Screenshots

1 / 1

Num

About this app

Num is a new AI-powered travel concierge that works like texting a knowledgeable friend. Instead of scrolling through countless reviews and travel apps, you simply tell Num what you want in any language, and it responds with the one place that fits — then books it for you. The service is free for travelers and covers over 2.7 million places across 106 destinations in 39 countries. What sets Num apart is its commitment to trust. Ratings come only from verified guests, and no business can pay to rank higher in results. Places that have claimed their listings are clearly marked, while the rest are simply mapped. Num speaks 16 languages around the clock, so whether you're asking for "dinner tonight, somewhere romantic I can walk to" or "best massage near me," you'll get a real answer in your own words. The app works two ways: you can add Num on LINE for a zero-install experience, or put the app on your home screen in just two taps without going through an app store. Bookings are held ...

Key features

1AI concierge that answers in natural language across 16 languages and books directly in the same thread
2Verified-guest-only ratings with a strict no-pay-to-rank policy — businesses cannot buy placement
3Two zero-friction entry paths: LINE bot (@799pyrus) for no-install use, or two-tap home-screen install without an app store

Target customer profile

Drawn from the product itself, not from a survey.

Demographic

International leisure travelers aged 25-45

Pain points

Overwhelmed by endless search results and unreliable reviews on traditional travel platforms

Primary needs

Quick, trustworthy local recommendations and instant booking without app-switching

Demographic

Business travelers in unfamiliar cities

Pain points

Limited time to research, language barriers, need for reliable same-day bookings

Primary needs

Fast concierge service in any language with confirmed reservations and schedule management

Demographic

Local hospitality businesses (restaurants, spas, tour operators)

Pain points

High commission fees, difficulty reaching travelers, time spent on back-and-forth booking coordination

Primary needs

Free listing with pay-per-completed-booking model and qualified guest leads

Scout Review

Written by AI from measured evidence, scored out of 100.

Num is an AI travel concierge that works like texting a friend — you ask in your own words, it answers with one place and books it. The pitch is sharp and the anti-pay-to-rank stance is a genuine differentiator against TripAdvisor, Google Maps, and Yelp, all of which let money influence what you see. The landing page is well-crafted and the two entry paths (LINE bot or two-tap home-screen install) remove real friction. Pricing is honest: free forever for travelers, $8.98/mo for Plus, $28.98/mo for Pro, with the FAQ explicitly stating that paying never buys better places. Where Num stumbles is infrastructure. Mobile performance is 69/100 with an 8.2-second LCP — bad for a product whose core use case is 'ask from your phone on the street.' Security headers score 2/12 with HTTPS not enforced, which is a real gap for a booking platform. Accessibility is 91/100 with three fixable failures. The idea is strong and the positioning is clear; the plumbing needs work before the concierge can be trusted with a credit card.

Measure by measure

Usability

Two low-friction entry paths (LINE or two-tap home-screen install), passwordless email sign-in, and a pricing FAQ that pre-empts real objections. Guides directory lacks visible search/filter in the fetched text.

Design

Clean, confident landing page with a live chat mockup that actually demonstrates the product instead of stock screenshots. Guides directory is dense but well-structured. No app-store UI to judge — the web surface is the product.

Speed

Desktop Lighthouse 94/100 is genuinely fast. Mobile is the problem: 69/100 with an 8.2s LCP — over three times the 'good' threshold. CLS 0 and TBT 137ms are clean, so it's a payload/render issue, not jank.

Security

Security-header probe scores 2/12: no enforced HTTPS, no HSTS, no CSP, no frame protection, no referrer or permissions policy. SPF and DMARC are present. No SOC 2, ISO, or bug-bounty signals found.

Accessibility

Lighthouse a11y 91/100 with three failing audits: prohibited ARIA attributes, insufficient color contrast, and non-sequential heading order. Ships a 10-language switcher and claims 16 languages. No published WCAG statement.

Growth

Real, specific problem with a structural differentiator (verified-guest-only ratings, no pay-to-rank) and an aligned pay-per-booking model. Early stage — domain is ~0.2 years old — but the wedge is legible.

Improvement areas

Security headers score 2/12 with HTTPS not enforced — critical gap for a booking platform handling user data
Mobile Lighthouse performance 69/100 with 8.2s LCP — unacceptable for a phone-first travel concierge
Guides directory (/guides/) lists 10,485 places across 227 guides with no visible search or filter UI

Design & Usability

Num's design does the one thing most travel sites fail at: it shows the product working. The homepage chat mockup — 'Just landed. Dinner tonight, somewhere romantic I can walk to?' answered in two blocks — communicates the concierge concept faster than any feature grid. Typography is disciplined, the ✦ motif is consistent, and the guides page at /guides/ presents 227 city directories with counted categories, which is dense but honest. Usability follows the same logic: two entry paths (LINE @799pyrus for zero-install, or a two-tap home-screen install that skips the app store entirely), passwordless email sign-in with a 15-minute link, and a pricing FAQ that answers the questions competitors dodge — 'Do I get better places if I pay?' gets a flat no. The weak spot is the guides directory: 10,485 places across 227 guides with no visible search or filter in the fetched content, which turns a useful resource into a scroll. Onboarding is genuinely fast; discovery inside the directory is not.

Speed & Security

Speed is a split verdict and the split matters. Desktop Lighthouse lands at 94/100 with CLS 0, TBT 137ms, and FCP 2.5s — that's a well-built page. Mobile lands at 69/100 with an LCP of 8.2 seconds, which is more than three times Google's 'good' threshold. Since Num's whole pitch is 'ask from your phone while walking to dinner,' the mobile number is the one that counts, and it's the one that's failing. No user complaints about lag surfaced in the collected evidence, so this reads as a payload/render issue rather than a systemic one — fixable, but currently real. Security is the sharper concern. The HTTP probe scores 2/12: HTTPS not enforced, no HSTS, no CSP, no frame protection, no X-Content-Type-Options, no Referrer-Policy, no Permissions-Policy. SPF and DMARC are present, and passwordless email-link auth is a reasonable session pattern, but a site that handles bookings and sends sign-in links without enforced HTTPS is leaving the front door propped. No SOC 2, ISO, or bug-bounty signals were found.

Accessibility & Growth

Accessibility sits at a solid-but-imperfect 91/100 on Lighthouse, with three concrete failures: prohibited ARIA attributes, insufficient color contrast, and non-sequential heading order. None are catastrophic; all are fixable in an afternoon. The multi-language story is the real strength — a 10-language switcher on the site and a product claim of 16 languages around the clock, which for a travel concierge is accessibility in the most literal sense. What's missing is a published WCAG conformance statement, which would turn 'we tried' into 'we committed.' Growth is a judgement on the idea, not the traction. Num names a real, specific problem: travelers drowning in review-scrolling and paid rankings, wanting one trustworthy answer. Its differentiator is structural, not cosmetic — verified-guest-only ratings, no pay-to-rank, and claimed-vs-mapped listing transparency. The business model (free for travelers, pay-per-completed-booking for hosts) aligns incentives in a way TripAdvisor's ad model structurally cannot. The honest limit: this wedge depends entirely on supply density and booking completion, and the domain is roughly 0.2 years old. The idea has room; the execution has a mountain to climb.

Conclusion

Num is trying to solve a problem everyone has felt: you're in a new city, you're hungry, and you're staring at 400 reviews that all say 'great food, slow service.' The answer — one trustworthy recommendation, booked in the same thread — is the right answer, and the no-pay-to-rank commitment is the kind of thing that only works if you actually mean it. The team appears to mean it. But a concierge is only as good as the house it works in, and right now the house has an unenforced HTTPS policy, no security headers worth mentioning, and a mobile page that takes eight seconds to load. None of that is unfixable, and none of it is unusual for a product this young. What matters is whether the next six months go into the booking flow or into the plumbing. The idea deserves the plumbing. Fix the headers, get the mobile LCP under three seconds, publish a security page, and Num has a real shot at being the thing people actually open when they land. Until then, it's a great pitch with a leaky roof.

Suggestions for improvement

Enforce HTTPS site-wide and add HSTS, CSP, X-Content-Type-Options, and Referrer-Policy headers — the 2/12 probe score is the biggest fixable gap.
Cut mobile LCP from 8.2s to under 2.5s: audit hero images, defer non-critical JS, and test on a throttled 4G profile, not desktop.
Add search and filtering to the /guides/ directory — 10,485 places across 227 guides is unusable without a way to narrow down.
Publish a security page and a WCAG conformance statement; both are cheap signals that build trust for a booking platform handling user data.

How we score

Alternatives comparison

Named competitors, point by point. Nobody paid to appear here or to be left out.

	Num	TripAdvisor	Google Maps / Google Travel	Yelp
Recommendation integrity (can businesses pay for placement?)	No — explicitly states no business can pay to rank higher; ratings only from verified guests	No — sponsored listings and paid placements influence visibility	No — rankings affected by ads and SEO rather than pure relevance	No — advertising packages affect business visibility and placement
AI conversational concierge (natural-language ask-and-book)	Yes — core product; ask in any of 16 languages, get one answer, book in the same thread	No — search-and-filter interface with AI trip planning features, not a conversational concierge	No — map-based search with Gemini integration, not a booking concierge	No — review and search platform with no AI concierge
In-thread booking and reservation management	Yes — books tables, tours, transfers; holds and confirms in the same conversation thread	Partial — bookings via partner integrations (Viator, TheFork) but not in a single thread	Partial — Reserve with Google for some restaurants; not a unified concierge thread	Partial — Yelp Reservations for restaurants; no tours, transfers, or unified thread
Global coverage and multi-language support	2.7M+ places across 106 destinations in 39 countries; 16 languages around the clock	Massive global coverage; supports many languages but not real-time conversational	Near-universal global coverage; strong multi-language support	Primarily North America and select markets; limited global coverage
Pricing model for travelers	Free forever for travelers; Plus $8.98/mo, Pro $28.98/mo for higher plan limits; hosts pay per completed booking	Free to use; revenue from ads, sponsored placements, and booking commissions	Free to use; revenue from ads and local business promotions	Free to use; revenue from advertising packages sold to businesses

TripAdvisor

A massive travel review and booking platform with millions of user-generated reviews, but recommendations can be influenced by paid placements and sponsored listings.

Google Maps / Google Travel

A ubiquitous mapping and local discovery tool with reviews and booking integrations, though rankings can be affected by ads and SEO rather than pure relevance.

Yelp

A popular local business review and discovery app, but primarily focused on North America and other select markets, with limited global coverage and no AI concierge.

Comparing options? See Num alternatives, scored side by side

View all Lifestyle & Entertainment