MCP Server
The Supp MCP (Model Context Protocol) server lets AI assistants interact with your support system directly. Classify messages, manage tickets, execute triggers, and monitor analytics — from your browser, terminal, or code editor.
Installation
You don't need a terminal or code editor. Supp works wherever you use AI — including Claude.ai in your browser. Authentication happens automatically via OAuth — no API keys to copy or manage.
Claude.ai (browser)
No terminal needed. In Claude.ai, go to Settings → Add custom connector. Set the name to supp and the URL to:
https://mcp.supp.support/mcpClick Connect to authenticate with your Supp account via OAuth. After that, you can manage your entire support system through conversation.
Claude Code (terminal)
Run this in your terminal:
claude mcp add --transport http supp https://mcp.supp.support/mcpThe first time you use an MCP tool, your browser will open to log in. After that, authentication is automatic.
Cursor
Add to .cursor/mcp.json in your project root:
{
"mcpServers": {
"supp": {
"url": "https://mcp.supp.support/mcp"
}
}
}Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"supp": {
"serverUrl": "https://mcp.supp.support/mcp"
}
}
}VS Code (Copilot)
Add to .vscode/mcp.json in your project root:
{
"servers": {
"supp": {
"type": "http",
"url": "https://mcp.supp.support/mcp"
}
}
}Other MCP Clients
Any MCP client that supports HTTP transport and OAuth can connect. Point it to:
https://mcp.supp.support/mcpYour browser will open to authenticate. No API keys or headers needed.
Available Tools
The MCP server exposes these tools to your AI assistant:
| Tool | Description |
|---|---|
| supp_help | Learn about Supp, available tools, and how to use them |
| supp_tickets | List classified support tickets with filtering |
| supp_ticket | Get full details of a single ticket |
| supp_ticket_resolve | Mark a ticket as resolved |
| supp_classify | Test classification against a message (no ticket created) |
| supp_intents | List all 315 supported intents with categories |
| supp_triggers | List all automation triggers |
| supp_trigger_create | Create a new trigger (intent to action mapping) |
| supp_trigger_update | Update an existing trigger |
| supp_trigger_delete | Delete a trigger |
| supp_trigger_simulate | Simulate a trigger against sample text (dry run, no side effects) |
| supp_queue | List pending approvals |
| supp_approve | Approve a pending action (executes immediately) |
| supp_reject | Reject a pending action |
| supp_integrations | List connected services |
| supp_connect | Get setup instructions for a service |
| supp_disconnect | Remove an integration |
| supp_stats | Overview metrics (conversations, resolution rate, costs) |
| supp_stats_intents | Intent breakdown with counts |
| supp_stats_triggers | Trigger performance metrics |
| supp_priority_score | Score a message's priority (low/medium/high/critical) |
| supp_classify_with_priority | Classify intent + score priority in one call |
| supp_priority_rules | List blanket priority rules |
| supp_priority_rule_set | Create or update a blanket priority rule |
| supp_priority_rule_delete | Delete a blanket priority rule |
| supp_routing_bulk | Bulk-update routing for multiple intents at once |
| supp_routing_category | Configure routing for an entire category |
| supp_routing_list | List all routing rules grouped by category |
| supp_categories | List all intent categories and counts |
| supp_escalated | List escalated tickets needing response |
| supp_escalation_respond | Respond to an escalated ticket |
| supp_balance | Check current balance and spend |
| supp_spend_cap | View or update spend cap settings |
| supp_docs_connect | Link a documentation source for templates |
| supp_trigger_execute | Execute integration actions FOR REAL (creates issues, sends messages) |
| supp_stats_export | Export support data as CSV or JSON |
| supp_refund | Submit a refund request for unauthorized charges |
| supp_refund_status | Check the status of refund requests |
Example Workflows
Set up bug report routing
Ask your AI assistant:
"Create a trigger that routes bug_report intents to Linear
with 0.85 confidence threshold, and also notify #bugs on Slack"Review and approve pending actions
Ask your AI assistant:
"Show me pending approvals and approve any refund
requests under $50 that have high confidence"Test classification
Ask your AI assistant:
"Classify this message: 'I was charged twice for my subscription'"Set up priority-based routing
Ask your AI assistant:
"Set up a blanket priority rule that pages on-call via Slack
for any critical message. Also configure all billing intents
to only fire their routing rules for high and critical priority."