Mireye

Mireye provides an infrastructure layer for AI agents to access data, enrichment, and tools that originate from the physical world. The product supports Claude, ChatGPT, Kimi, Gemini, Cursor, Grok, Copilot, DeepSeek, or custom agents through a single API and MCP tools. Output is returned with citations, provenance and access to a catalog of fields and signals. The system is designed to respond to natural-language questions with data pulled from authoritative sources and a clear record of where each value comes from.

It acts as the physical-world context layer for AI agents, delivering data, enrichment, and tools that agents can leverage to reason about real places. The interface centralizes data requests, field calculations, and signal outputs, while ensuring traceability to original sources and fetch timestamps for each value.

The architecture emphasizes a catalog of fields and presets that guide retrieval, enrichment, and calculations. The catalog is extensible, supporting requests for new fields and matching gaps to queued builds that become available to all callers after completion. Output includes a plan of the executed steps so agents can replay decisions.

Use cases

• Fetch public coordinate data or enrich parcels for site-siting analyses in real-world contexts.
• Find parcels by requirements and compare candidates with enrichment and citations.