Open Source · by Joan Sanz

Hospitality
AI Agents
Repository

18 production-ready AI agent specs for hotels — check-in, revenue, operations, training and more. Model-agnostic. Runs on any LLM. Free forever.

View Repository Generate My Agents The Future
18
AI Agent Specs
5
Departments
4+
LLMs Compatible
MIT
License · Free
The three tools

Everything you need to
deploy AI in your hotel

Three standalone tools. Each one works independently. Together they take you from zero to a fully deployed AI agent stack.

01
Repository
📚
Hospitality AI Agents
18 complete agent specifications — system prompts, KPIs, escalation rules, and implementation guides for every hotel department.
Check-in, Concierge, Complaints, Upsell, Loyalty
Dynamic Pricing, Overbooking, Forecasting
Training simulators for hospitality schools
GDPR compliance + safety guardrails included
Open Repository
18 agents
02
Generator
⚡
Agent Generator
Fill in your hotel's details and generate a personalized .zip package — all agent specs tailored with your hotel's name, PMS, tone and language.
Personalized with your hotel name, city & brand voice
Choose your PMS, WhatsApp provider and language
Live preview before download
Download a ready-to-use .zip — no signup required
Open Generator
No signup
03
Vision
🔭
The Agentic Future
Why every PMS, booking engine and channel manager has no brain today — and what the adoption roadmap looks like from 2025 to full autonomy.
The technology gap the industry isn't talking about
Claude vs GPT-4o vs Gemini for hospitality
Can an AI agent actually log in and do things?
Industry adoption curve 2025 → 2027+
Read the Vision
Full manifesto
What is this

A Markdown file
is the agent's brain

Each .md file in this repository is a complete agent specification. Load it into any LLM as a system prompt and you instantly have a trained hospitality professional. The spec is portable, version-controlled, and works with Claude, GPT-4, Gemini, or any future model.

📄
Write once, deploy anywhere
The same .md spec works with Claude, GPT-4o, Gemini or Llama. Switch models without rewriting your intelligence.
🔁
SOP that thinks
Standard Operating Procedures that an AI can actually read, reason about, and execute — not just a PDF in a drawer.
🏨
Built for real hotels
Every spec includes system prompt, KPIs, escalation rules, compensation limits and PMS integration notes.
checkin-agent.md → Claude Opus
# Check-in Agent — Hotel Example ROLE: Welcome & arrival assistant HOTEL: Hotel Example, Barcelona TONE: Warm, professional, efficient --- TRIGGER: Guest message detected → get_reservation(guest_id) → check_loyalty_tier() → get_room_status() DECISION: if tier == "Platinum" → activate_vip_protocol() → trigger_suite_upgrade() else → send_standard_welcome() ESCALATE if: payment_issue | complaint KPI target: <3min check-in | >4.5★ █
How it works

From .md spec
to autonomous hotel

Four steps. No code required for step 1. Full API autonomy by step 4.

1
📋
Choose your agents
Select the agents relevant to your hotel — check-in, complaints, pricing. Download the .md specs or generate a personalized package.
2
🔌
Load into your LLM
Paste the spec as a system prompt in Claude, GPT-4, HiJiffy or Asksuite. You now have a trained hospitality AI. No fine-tuning needed.
3
📱
Connect to WhatsApp
Link your LLM to WhatsApp Business API via Twilio or MessageBird. Your agent is now live — responding to guests in real time.
4
🔗
Connect your PMS
Add API access to Mews, Opera or Cloudbeds. Your agent can now read reservations, trigger tasks and adjust rates autonomously.
18 agents

Every department. Every role.

From the moment a guest books to the review they leave after checkout.

Check-in AgentP1
Complaints AgentP1
Concierge AgentP2
Upsell AgentP2
Loyalty AgentP2
Pricing AgentP2
Overbooking AgentP2
Forecast AgentP3
Housekeeping AgentP3
Maintenance AgentP3
F&B AgentP3
Review AgentP2
Email CampaignsP2
B2B OutreachP3
Groups & EventsP3
Onboarding AgentSchool
Roleplay TrainerSchool
QA AgentSchool
Guest Experience
Revenue Management
Operations
Marketing
Training
P1=Phase 1 (week 1-4) · P2=Phase 2 · P3=Phase 3
// The meta vision
"When any LLM can consume a .md file and act as a trained hospitality professional —
the repository that defines those specs becomes the intelligence layer the entire industry runs on."
— Joan Sanz, Hospitality AI Agents Repository
Read the full vision
Adoption curve

The industry roadmap
to full agent autonomy

Now — 2025
The Spec Layer
Copy .md into Claude or GPT-4 as a system prompt. Immediate intelligence — works today.
2025–2026
Platform Bridge
HiJiffy, Asksuite, Quicktext add LLM APIs. These specs become their source of truth.
2026–2027
PMS Goes Agentic
Mews, Opera, Cloudbeds build AI layers. The .md spec is their hotel config file.
2027+
Full Autonomy
The agent has credentials for every system. Reads arrivals, adjusts rates, dispatches housekeeping — all from a .md config.
// Open source · free forever
Start this week.
Be two years ahead.

Download the repository. Generate your personalized agent pack. Run your first agent via WhatsApp. No vendor contract. No signup. Just the spec.

View Repository Generate My Agents The Agentic Future