DealFlowPro MCP Server is an MCP (Model Context Protocol) server provided by the DealFlowPro platform, used to analyze multifamily real estate deals from Claude Code, Cursor, Claude Desktop, or any MCP-compatible client. DealFlowPro itself is an AI-driven multifamily deal analysis platform, and this MCP server wraps the DealFlowPro REST API, enabling AI agents to access institutional-grade underwriting calculations.
Purpose Users can ask Claude questions in natural language, and Claude will automatically call the correct tools. For example: analyze a 24-unit deal in Charlotte (asking price $2 million, monthly rent $13,600, expenses $5,500); score a 12-unit, $1.5 million deal with monthly income of $9,000; what is the maximum price to offer for a property with monthly rent of $28,800 if you want an 8% cash-on-cash return; check whether an address is in a flood zone.
Main capabilities The four calculation tools all carry MCP 2025-06-18 annotations (title, readOnlyHint, destructiveHint, idempotentHint, openWorldHint), and each tool is read-only. When installed as a Claude plugin, it connects to the remote server https://dealflowpro.io/mcp，而非 npm package, so it provides eight tools instead of four: it additionally adds list_my_deals, get_deal_analysis, get_my_criteria, and portfolio_summary, used to access saved deals, buy-box criteria, and pipeline, scoped to the user account. The plugin uses OAuth instead of an API Key, so there is no need to export DFP_API_KEY; simply log in to your DealFlowPro account when connecting. If you need a local stdio server, prefer API Key authentication, or want to connect DealFlowPro to a non-Claude MCP client, you should use the npm package.
For institutional AI teams DealFlowPro can be integrated into AI evaluation pipelines and production AI stacks. The MCP server (npm package and remote endpoint) wraps the same engine as the REST API and is characterized by strict schemas, idempotency, and predictability. In terms of authentication and scoping, every tool call is authenticated (Bearer API Key or OAuth login for the remote endpoint), calls are scoped to the account that issued the key or logged in, and there is no cross-tenant access path. An audit line for each tool is recorded in the account's logs/mcp_tool_calls.log, capturing the tool name and flattened parameter keys (without values), so usage is auditable without exposing deal contents. Rate limits are divided by tier, each response returns rate limit headers (X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset), and 429 responses include Retry-After. Error codes are stable and documented at dealflowpro.io/api/docs/#errors, and all errors are in the JSON format {success: false, error: {code, message}}, with codes including invalid_input (400), invalid_json (400), unauthorized (401), tier_required (403), method_not_allowed (405), payload_too_large (413), rate_limit_exceeded (429) internal_error (500). Regarding the tool schema, each MCP tool has a strictly defined input schema, and the runtime canonical schema can be obtained by calling tools/list on the remote endpoint; the REST OpenAPI specification covers the same shapes for the four calculation tools (analyze_deal, score_deal, reverse_calc, market_data). The four saved pipeline tools on the remote server (list_my_deals, get_deal_analysis, get_my_criteria, portfolio_summary) are MCP-only and require an Essentials plan or higher. Evaluation framework pattern recommendation: maintain a fixed test set of (deal payload → expected metrics), run it against score_deal (the lightest tool, each call still counts as one request), and assert that metrics are within tolerance; the endpoint is deterministic, with identical inputs producing identical outputs. Regarding data handling, request payloads are processed in memory and are not persisted to disk; transport uses TLS 1.2+; when DealFlowPro internally uses Claude for document extraction, Anthropic API calls go through DealFlowPro's zero data retention Anthropic workspace. The full security posture is at dealflowpro.io/security#api-mcp-data-handling.
Usage There are two installation paths. Path A — remote MCP URL (no local installation required): suitable for claude.ai web (custom connector), Claude Code using HTTP transport, or the Claude API mcp_servers. The endpoint is https://dealflowpro.io/mcp（Streamable HTTP, MCP 2025-06-18). In claude.ai web / Claude Desktop, DealFlowPro is already listed in the Claude Connectors Directory and can be added directly, or you can add it yourself as a custom connector: Customize → Connectors → "Add custom connector" → name it DealFlowPro, paste https://dealflowpro.io/mcp → Add → Connect → log in to DealFlowPro → click Allow. OAuth handles authentication automatically (no API key management required). Then in chat, click "+" → Connectors → turn on the DealFlowPro toggle. Members of Team/Enterprise plans cannot add custom connectors themselves; an organization Owner must add it once in Organization settings → Connectors → Add → Custom (type: Web), and then each member connects their own DealFlowPro account from Customize → Connectors. Claude Code (HTTP) requires obtaining a Bearer key at dealflowpro.io/api (a free 5-request key, no credit card required, or pay-as-you-go credits). Full setup details and connected app revocation are at dealflowpro.io/api/docs#mcp. Path B — local stdio (this npm package): suitable for cases where you want to run the MCP server locally or the client has no remote URL path. Claude Desktop supports both: use remote via Settings → Connectors, or use the local configuration below. First obtain an API key at dealflowpro.io/api (a free 5-request key, no credit card required, or purchase pay-as-you-go credits), sent via email. Then add it to the client, for example Claude Desktop's claude_desktop_config.json. After that, simply ask Claude about deals, and it will automatically select the correct tool.
When to use it Suitable for operators, syndicators, and private equity firms to quickly screen deals. DealFlowPro automates multifamily underwriting for 2-200 unit properties; upload a broker email or OM and get a complete investment analysis in 10 minutes instead of 2 hours. It is also suitable for enterprise AI teams that need to connect DealFlowPro into AI evaluation pipelines and production AI stacks.
Limitations The four calculation tools are read-only; the four saved pipeline tools on the remote server require an Essentials plan or higher. Rate limits are tiered by plan. On pricing: pay-as-you-go $1/API request, starting at $5; Essentials is $79 per month, including 50 API requests per day; Premium is $149 per month, including 200 requests per day and a reverse calculator; Enterprise is $399 per month, including 1,000 requests per day. Request payloads are processed in memory and are not persisted to disk.
