▸ hotel research agent
live · monad testnetletmefind
Describe the hotel you want. An AI agent finds it.
Tell us destination, budget, and must-haves in plain English. We search live listings, score the best matches, and show why each stay fits — so you skip hours of tab-hopping.
delegated search,
not another chatbot.
LetMeFind turns a plain-English request into personalized hotel recommendations. It completes the research end to end — then explains the picks.
// mvp: hotel discovery · research agent · on-chain memory
▸ the problem
the research loop is still manual.
Multiple sites, repeated filters, price comparisons, reviews, amenities — users still do most of the work, especially when they care about strict budgets, reliable WiFi, family-friendly stays, business travel, accessibility, proximity to attractions, and flexible cancellation.
you know what you want. letmefind finds it.
▸ the ask
“Find me a beachfront hotel in Bali for under $500 with breakfast, excellent WiFi, and a swimming pool.”
The agent understands the request, researches sources, ranks by your priorities, and returns the strongest matches with reasons — no filter forms across booking sites.
▸ pipeline
three agents. one pass.
planner → search → ranking · each has one job
01
[planner]
Turns plain language into structured criteria — destination, budget, dates, amenities, and special requirements.
02
[search]
Gathers listings, ratings, amenities, and prices; dedupes and normalizes. Collects facts — does not pick a winner.
03
[ranking]
Scores fit against your request and writes a clear explanation — budget, reviews, amenities, location, value.
▸ results
concierge, not a browse page.
A short assistant reply, then an equal grid of recommendation cards — photo, match %, why bullets, and booking CTAs. Refine in conversation; no filter chrome.
▸ sample reply
“I researched Bali under $450 with breakfast. Here are stays that fit — each card shows match score and why I’d recommend it.”
▸ input
talk like a traveler.
“I need a hotel in Paris under $350 with great WiFi because I’ll be working remotely.”
understand → extract → research → compare → rank → explain → present
▸ access
search credits in mon
Connect a Monad Testnet wallet. Exchange native MON for prepaid credits — access to the agent, not a booking fee.
- credit pack
- 0.5 MON → 10
- each find
- 1 credit
- testnet mon
- faucet.monad.xyz ↗
▸ ownership
on-chain research records
Optionally save completed research under your wallet — destination, budget, recommended hotel, confidence, timestamp, and fingerprint. Not a booking ledger: a receipt of the agent’s work.
▸ principles
how we ship
- natural language first
- Describe what you need. No filter labyrinth.
- autonomous research
- The agent compares listings so you don’t have to.
- explainable picks
- Every recommendation says why it won.
- user ownership
- Research can live on-chain under your wallet.
- modular agents
- Same plan → research → rank path for future domains.
hotels first. same pipeline later.
Flights, apartments, restaurants, activities, itineraries, visa info — each can plug into plan → research → recommend without rewriting the product.
▸ open
delegate the research. keep the record.
Connect, exchange MON for credits, describe the stay — transparent recommendations you can save on Monad Testnet.
→ ask the concierge