MCP Server Developer — Real Estate AI Integration (ARMLS Spark API + Claude)

:I’m the owner of a real estate brokerage in Maricopa, Arizona and I’m building an internal AI platform for my agents using Claude (by Anthropic). I need a developer to build a small MCP (Model Context Protocol) server that connects our ARMLS MLS data feed to Claude so agents can query live listing data directly in conversation. What MCP is: MCP is Anthropic’s open protocol for connecting Claude to external data sources. The server exposes tools that Claude can call during a conversation — think of it as a bridge between Claude and our MLS data. Documentation is at modelcontextprotocol.io. What needs to be built: A lightweight MCP server that authenticates with the ARMLS Spark API (FlexMLS Web API) and exposes the following tools to Claude: 1. Search active listings by city, zip, price range, property type, status 2. Pull comparable sales for a specific address with radius and date range 3. Get market statistics by zip or city — median price, DOM, list-to-sale ratio, inventory count 4. Look up a specific listing by MLS number 5. Pull active vs. sold trend data by area over a time range 6. Open house data by area and date range (nice to have) Technical requirements: ∙ Node.js or Python — your choice ∙ MCP protocol compatible with Claude.ai Team connectors ∙ HTTP/SSE transport (remote server, not local) ∙ Authenticates with ARMLS Spark API credentials I will supply ∙ Hosted on Railway, Render, or similar — must be publicly accessible via HTTPS ∙ Credentials stored as environment variables, never hardcoded ∙ README with setup steps, environment variables, and instructions for adding new tools ∙ GitHub repo (private) — full ownership transfers to me About the ARMLS Spark API: ARMLS uses the Spark Platform (FlexMLS Web API). Documentation at sparkplatform.com. I have an existing IDX Content License Agreement with ARMLS and am in the process of completing a new data licensing agreement that will authorize this internal back-office use. The developer will need to be listed as a Consultant on the ARMLS Content License Agreement — I handle that process, I just need your name and company. This is strictly internal: No public IDX display. No consumer-facing website. Internal back-office tool for agent use only. Deliverables: ∙ Working MCP server deployed and accessible via HTTPS ∙ All tools functional and tested ∙ README documentation ∙ Private GitHub repo with full code ownership transferred ∙ Brief Loom walkthrough showing it working end to end Budget: Fixed price — $300 to $600 depending on experience and scope Timeline: 2 weeks or less preferred To apply, please include: ∙ A specific example of an API integration or MCP server you have built ∙ Your proposed tech stack and hosting platform ∙ Your fixed price bid and estimated timeline ∙ Confirmation that you have reviewed the Anthropic MCP documentation at modelcontextprotocol.io ∙ Any questions about the ARMLS Spark API Developers who demonstrate specific MCP knowledge in their proposal will be prioritized. Generic proposals will not be considered. Job Settings to select on Upwork: ∙ Category: Web Development → API Development ∙ Skills to add: API Integration, Node.js, Python, REST API, MCP ∙ Project type: Fixed Price ∙ Budget: $300–$600 ∙ Duration: Less than 1 month ∙ Experience level: Intermediate to Expert