Routestackai

The Travel Infrastructure Layer for AI Agents

GitHub

RouteStack.ai — Starter Templates

Boilerplate projects for connecting to the RouteStack MCP server. Each starter is standalone — clone it, add your API key, and run.

Quick Start

cd typescript/cli-agent
cp .env.example .env
# Add your ROUTESTACK_API_KEY
pnpm install
pnpm start

Available Starters

TypeScript

StarterDescription
chat-agentBrowser-based chat UI for travel queries
cli-agentTerminal-based travel search tool
slack-botSlack slash commands for travel
discord-botDiscord bot for travel search
react-widgetEmbeddable <TravelSearch /> component
multi-agentMulti-agent trip planner (Claude Agent SDK)
webhook-listenerSSE event subscriber + forwarder
nextjs-ai-mcpNext.js + Vercel AI SDK integration
voice-agentVoice AI travel agent (OpenAI Realtime / Vapi)
ollama-localLocal LLM (Ollama) + remote RouteStack MCP
n8n-mcp-nodeCustom n8n node for travel automation
chrome-extensionContext-aware Chrome side panel agent
price-drop-cronCloudflare Worker cron for price alerts
corp-approval-flowCorporate travel approval via Deep Links
rag-travel-profilePersonalized search with vector DB

Python / Go / Rust

Coming soon. See the language matrix for planned coverage.

Prerequisites

  • A RouteStack API key (get one at routestack.ai)
  • Node.js >= 20 (for TypeScript starters)
  • An LLM API key (OpenAI or Anthropic) for agent-based starters

Configuration

Every starter uses the same core environment variables:

ROUTESTACK_API_KEY=your_api_key_here
ROUTESTACK_MCP_URL=https://mcp.routestack.ai/sse

Some starters require additional keys (LLM provider, Slack, Discord, etc.). See each starter's .env.example for details.

相關伺服器