QCX

Language to Maps

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

Command

claude mcp add --transport stdio queuelab-qcx node path/to/server.js \
  --env EXA_API_KEY="" \
  --env XAI_API_KEY="https://console.x.ai" \
  --env GOOGLE_API_KEY="optional (Gemini)" \
  --env OPENAI_API_KEY="optional" \
  --env TAVILY_API_KEY="" \
  --env BEDROCK_ACCESS_KEY_ID="optional" \
  --env UPSTASH_REDIS_REST_URL="" \
  --env UPSTASH_REDIS_REST_TOKEN="" \
  --env BEDROCK_SECRET_ACCESS_KEY="optional" \
  --env NEXT_PUBLIC_MAPBOX_ACCESS_TOKEN=""

How to use

QCX is an experimental AI-first geospatial companion that lets you explore the planet and beyond through natural language and interactive maps. You can chat with an AI about locations, events, or spatial questions, then draw on the map to obtain measurements, analyses, and AI-assisted insights. The system supports multi-model backends and a tool-using agent workflow to perform geospatial queries, fetch data from maps, and render results in a generative UI. The experience emphasizes a bidirectional integration between chat and map interactions, enabling real-time reasoning with maps, 3D views, and time-aware contexts. To get started, run the server locally or deploy the app, then connect with the UI to begin a conversational exploration of places, routes, and geographic phenomena. The toolchain is designed to minimize unnecessary LLM calls while routing tasks to appropriate geospatial tools and models (Grok, OpenAI, Google, Bedrock, etc.).

How to install

Prerequisites:

Node.js or Bun installed (this project uses Bun in practice, but Node.js compatibility checks are supported).
Git client to clone the repository.

Installation steps:

Clone the repository:

git clone https://github.com/QueueLab/QCX.git
cd QCX

Install dependencies (Bun is recommended as per the project instructions):

bun install

Configure environment variables:

cp .env.local.example .env.local

Edit .env.local to provide API keys and tokens as needed (XAI_API_KEY, OPENAI_API_KEY, GOOGLE_API_KEY, BEDROCK_*, TAVILY_API_KEY or EXA_API_KEY, UPSTASH_REDIS_REST_URL/Token, NEXT_PUBLIC_MAPBOX_ACCESS_TOKEN).

4) Run the development server:

bun run dev

Or for production preview:

bun run build && bun run start

Open http://localhost:3000 to access the QCX interface.

Additional notes

Notes and tips:

Ensure at least one AI provider key is configured (XAI_API_KEY, and optionally OPENAI_API_KEY or other providers) for tool-using agents to operate.
Map functionality relies on NEXT_PUBLIC_MAPBOX_ACCESS_TOKEN; without it, map rendering will be unavailable.
If you encounter formatting issues with model outputs, consider adjusting the model selection (Grok, OpenAI, Google, Bedrock) or enabling frontier models for better tool integration.
Redis (UPSTASH_REDIS_REST_URL and UPSTASH_REDIS_REST_TOKEN) can improve rate limiting and prompt storage performance.
For local development, Bun is recommended; if using Node.js, ensure compatibility with your environment.
Review the in-depth architecture and component docs at the provided deepwiki link for advanced customization and tooling options.