Skip to content

KamraThe hotel PMS agents can actually run

Open-source PMS with a governed MCP layer - Claude (or any agent) quotes, books and posts charges as a permission-checked user. Prices stay in code. Every action is audited.

Kamra

What makes Kamra different ​

Most hotel software bolts a chatbot onto a closed SaaS. Kamra is built so an agent and a receptionist share one governed write path.

  1. MCP first. Connect Claude (or any MCP client) to /mcp. Tools are filtered by the signed-in user's roles and the property's enabled modules.
  2. Deterministic money. Quotes, taxes, availability and cancellation fees come from pricing.py and the folio ledger - never from the model. Guardrails refuse out-of-range rates.
  3. Full audit. Every mutating tool call lands in the Activity Log with who / what / why.
  4. Same APIs for everyone. Desk UI, REST, MCP and the in-app Kamra Agent call the same @frappe.whitelist functions. No second writer.

Start here: Connect your AI (MCP) · MCP tool reference · How Kamra works · Developer shortcuts

Current stable: v2.6.17 · Live demo · Frappe Cloud Marketplace

Kamra's Today screen: arrivals, departures, the room board and in-house guests

Where to start ​

Quick links: Connect MCP · MCP tools · API basics · API reference · Postman ↓ · What's new · changelog

Open source (AGPL-3.0) · docs for Kamra v2.6.17 · every feature included, always.