How does ARD relate to MCP?

ARD is the discovery layer, MCP is the call layer. An agent finds your MCP server URL via ai-catalog.json, then calls its tools via MCP. The two are complementary, not competing.

ARD reached broad hyperscaler backing in July 2026 with Google, Microsoft, Salesforce, Snowflake, ServiceNow, Cisco, Databricks, GitHub, NVIDIA, Hugging Face, and GoDaddy. GitHub Copilot Agent Finder and Hugging Face Discover Tool are already implementing it.

Why should a hotel platform publish ai-catalog.json?

It is the fastest way to become agent-discoverable, and no meaningful hotel or travel platform has one live yet. Zero-cost implementation (a static JSON file), zero competition in the category, and it compounds as agent frameworks adopt ARD-based discovery.

What data does Vervotech expose for ARD?

3M+ canonical properties across 400+ suppliers at 99.998% mapping accuracy, available via REST and stateless MCP endpoints. Every endpoint is ready to be listed in an ai-catalog.json entry — canonical property IDs, room mappings, and static hotel data.

Does implementing ARD cost anything?

The ai-catalog.json file itself is a static JSON file — free to ship. The upstream data substrate that makes it valuable (canonical hotel IDs, MCP endpoints, mapped supplier catalog) starts at $499 a month with Vervotech Hotel Mapping.

Agentic Resource Discovery for Hotels

Agentic Resource Discovery Hotels is the open discovery standard that lets AI agents find and call structured data at your domain. Sites publish /.well-known/ai-catalog.json listing their MCP servers, OpenAPI specs, and knowledge entries so agents can auto-discover capabilities without a human integrator in the loop.

For hotel and travel platforms, ARD is the fastest path to being agent-discoverable. Vervotech ships the substrate: 3M+ canonical properties across 400+ suppliers at 99.998% accuracy, exposed via REST and MCP, ready to be listed in an ai-catalog.json.

Why ARD crossed the industry threshold in July 2026

The Agentic Resource Discovery specification (v0.9) reached broad hyperscaler backing in July 2026, positioning it alongside MCP as a foundational agent standard.

11 hyperscalers backing ARD: Google, Microsoft, Salesforce, Snowflake, ServiceNow, Cisco, Databricks, GitHub, NVIDIA, Hugging Face, and GoDaddy.
GitHub Copilot Agent Finder and Hugging Face Discover Tool are already implementing ARD-based discovery.
Federation model: ai-catalog.json points to your MCP servers and APIs; agents follow the trail with zero pre-configured integrations.
Complementary to MCP: ARD handles discovery, MCP handles the call. Both are needed.

Hotel and travel platforms have effectively zero competition in the ARD space right now — a rare first-mover window.

What belongs in a hotel platform ai-catalog.json

mcp_servers: One or more MCP server URLs. Vervotech ships a stateless MCP server exposing search_hotels, get_property, get_rooms, and resolve_duplicates.
openapi_specs: Machine-readable REST API descriptions so agents can call endpoints without human docs.
knowledge_entries: Structured citations to canonical documentation: mapping methodology, freshness policy, accuracy benchmarks.
contact: Programmatic contact endpoints for agents that need to escalate or clarify.
auth: Documented auth flows — API keys, OAuth, or open discovery for public endpoints.

What shipping ARD unlocks

Auto-discoverability from GitHub Copilot, Hugging Face, and any ARD-compliant agent framework.
Zero integration friction for downstream partners — new agents find you at /.well-known/ai-catalog.json and start calling.
First-mover positioning in the hotel and travel category, where no meaningful competitor has an ai-catalog.json live.
Compounding trust as agents build usage graphs around discoverable endpoints.

Vervotech is the mapped-data substrate for hotel ARD

3M+ canonical properties across 400+ suppliers at 99.998% accuracy, exposed via REST and MCP endpoints ready to be listed in your ai-catalog.json. Free 7-day trial.

Product Tour

Frequently asked questions

What is Agentic Resource Discovery (ARD)?: ARD is an open discovery standard backed by 11 hyperscalers in 2026. Sites publish /.well-known/ai-catalog.json listing their MCP servers, OpenAPI specs, and knowledge entries. AI agents fetch that file and auto-discover what a domain can do, without a pre-configured integration.
How does ARD relate to MCP?: ARD is the discovery layer, MCP is the call layer. An agent finds your MCP server URL via ai-catalog.json, then calls its tools via MCP. The two are complementary, not competing.
Who is backing ARD?: ARD reached broad hyperscaler backing in July 2026 with Google, Microsoft, Salesforce, Snowflake, ServiceNow, Cisco, Databricks, GitHub, NVIDIA, Hugging Face, and GoDaddy. GitHub Copilot Agent Finder and Hugging Face Discover Tool are already implementing it.
Why should a hotel platform publish ai-catalog.json?: It is the fastest way to become agent-discoverable, and no meaningful hotel or travel platform has one live yet. Zero-cost implementation (a static JSON file), zero competition in the category, and it compounds as agent frameworks adopt ARD-based discovery.
What data does Vervotech expose for ARD?: 3M+ canonical properties across 400+ suppliers at 99.998% mapping accuracy, available via REST and stateless MCP endpoints. Every endpoint is ready to be listed in an ai-catalog.json entry — canonical property IDs, room mappings, and static hotel data.
Does implementing ARD cost anything?: The ai-catalog.json file itself is a static JSON file — free to ship. The upstream data substrate that makes it valuable (canonical hotel IDs, MCP endpoints, mapped supplier catalog) starts at $499 a month with Vervotech Hotel Mapping.