MCP Agents

Agent LayerUse MCP agentswhere language improves the workflow

TripGuru MCP agents sit on top of the same API primitives and help with discovery, operations, support, and internal analysis.

REST-backedConversational accessOperationally constrained

Best fit

Good for: search, support triage, reporting, and guided operational tasks.

Not the first step: core transactional reliability still comes before agent orchestration.

Agent Lineup

Product search agent

Turns broad natural-language requests into filtered product discovery.

Intent parsingCatalog searchRecommendation framing

Booking operations agent

Moves from quote context into cart, order, and booking changes.

Cart updatesOrder lookupOperational actions

Support agent

Handles order status, refund context, and common traveler questions.

Order answersRefund contextEscalation prep

Analytics agent

Summarizes what happened across bookings, channels, and time windows.

Sales summariesTrend promptsOperational reporting

Config

opencode.json

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "tripguru-mcp": {
      "type": "remote",
      "url": "https://tripguru.ai/mcp",
      "enabled": true,
      "oauth": false,
      "timeout": 30000,
      "headers": {
        "Authorization": "Bearer <your-api-key>"
      }
    }
  }
}

Typical Prompts

“Find family-friendly tours in Paris under €100 per traveler.”

“Change tomorrow’s booking to next Tuesday and confirm the price delta.”

“Summarize all pending refunds created in the last seven days.”

“List the highest-converting product categories from the previous month.”

Rollout

The safest agent rollout is narrow, observable, and intentionally boring.

Start with a workflow that benefits from natural language but does not hide critical system transitions.

Stabilize the API flow

Make sure product, cart, order, and payment paths already work without an agent in the loop.

Expose the MCP server

Connect your assistant runtime to the TripGuru MCP endpoint with the correct environment variables.

Constrain the use case

Pick a narrow workflow first, such as product search or order support, and define what the agent may modify.