Model Context Protocol
External AI Agents
SendSquared exposes a Model Context Protocol server, so your team can connect any LLM — Claude, ChatGPT, Grok, Gemini, and more — to your CRM. Ask questions in plain language, surface insights, and trigger workflows. Every model amplified by your hospitality data.
Compatible Models
Every LLM, Squared by SendSquared
Claude Anthropic via MCP 
ChatGPT OpenAI via MCP 
Grok xAI via MCP 
Gemini Google via MCP 
Llama Meta via MCP 
Mistral Mistral AI via MCP 
Perplexity Perplexity via MCP 
DeepSeek DeepSeek via MCP Claude Anthropic via MCP ChatGPT OpenAI via MCP Grok xAI via MCP Gemini Google via MCP Llama Meta via MCP Mistral Mistral AI via MCP Perplexity Perplexity via MCP DeepSeek DeepSeek via MCP Ask Your CRM Anything
Connect Claude, ChatGPT, or any MCP-compatible client to SendSquared and ask questions in plain language. Pull occupancy, lifetime value, channel performance, agent productivity, and call analytics without writing reports or exporting data.
Natural-Language Queries
Ask "how is direct booking pacing this week?" and get a real answer pulled from live CRM data.
Live Analytics
Surface revenue attribution, occupancy, and campaign performance without dashboards.
Guest Profile Lookups
Find guests by name, phone, or stay history and pull their full profile.
Auditable Reads
Every query logged with the agent, model, and tenant for compliance.
Trigger Workflows from a Conversation
With elevated scope, agents can act on the CRM — create leads, send messages, trigger automations, update guest records. Tell Claude or ChatGPT what you want done in plain language; SendSquared executes it through the same APIs your team uses, with the same permissions.
Trigger Automations
Fire campaigns, win-back sequences, or pre-arrival flows from a conversation.
Send Messages
Compose and send email or SMS to specific guests or segments.
Create & Update Records
Create leads, update guest profiles, and tag contacts based on AI judgment.
Scoped Permissions
Match the same ACL controls your team uses — read-only or write, narrow or broad.
Build Custom Agents on Top of SendSquared
Your engineering team can build custom AI agents that pull SendSquared data and act on it — internal copilots, analyst sidekicks, voice-driven dashboards. The MCP server is the foundation; the agents are yours to design.
Custom Copilots
Build internal AI tools that surface CRM data on demand.
Open Standard
MCP is an open protocol — no vendor lock-in to a single AI provider.
Multi-Tenant Ready
Token-scoped per property or brand for multi-property operations.
Automatic Updates
New SendSquared APIs become available to agents without code changes.
Security and Audit Trail
MCP access uses tenant-scoped tokens. Every query is logged with the model, agent, and timestamp. Write actions can be restricted to specific scopes (read-only, send-only, full). Tokens can be revoked or rotated instantly. Your data, your permissions, your audit trail.
Token-Scoped Access
Per-tenant, per-property, per-scope tokens with instant revocation.
Full Audit Trail
Every read and write logged with model, agent, and timestamp.
Permission Match
MCP scopes mirror your existing ACL controls — same rules, same enforcement.
Token Rotation
Schedule automatic token rotation or revoke at any time.
Operators querying through MCP can ask agents for portfolio-wide rollups — including multi-property lifetime value — and get answers pulled from live CRM data without exporting a single report.
Frequently Asked Questions
What is MCP?
Model Context Protocol is an open standard that lets AI assistants like Claude and ChatGPT securely connect to external systems. SendSquared runs an MCP server, so any MCP-compatible agent can query your CRM, surface insights, and trigger workflows on your behalf.
Which AI models can I use?
Any model that supports MCP. Today that includes Claude, ChatGPT, custom agent frameworks, and emerging support across Grok, Gemini, Llama, Mistral, Perplexity, and DeepSeek. As MCP adoption grows, every new model becomes immediately compatible.
What can the AI actually do with my data?
Read-only by default: surface guest profiles, recent reservations, lifetime value, campaign performance, agent productivity, and call analytics. With elevated scope, agents can also create leads, send messages, trigger automations, and update guest records — every write action is logged with full audit trail.
Is it secure?
Yes. MCP access is gated by tenant-scoped tokens, granular permissions match your existing ACL controls, every query is logged, and write actions can be limited to specific scopes. Tokens can be revoked instantly and rotated on a schedule.
Who is this for?
Operators and analysts who want to query the CRM in plain language ("how is direct booking pacing this week?"), and ops teams who want AI-driven workflows ("send a follow-up to every guest who arrived last week and rated their stay below 8"). It also opens the door to custom AI agents built specifically for your operation.
How do I set it up?
MCP access is available on enterprise plans. Setup takes about 30 minutes — generate an MCP token in SendSquared, add the server URL to your AI client (Claude Desktop, ChatGPT custom agents, or any MCP-compatible tool), and start asking questions.
Connect Your AI Stack to SendSquared
Book a demo to see MCP in action — Claude, ChatGPT, or any agent of your choice running queries against a live SendSquared environment.
Schedule a Demo