How FollowUpAce's New MCP Integration Turns Claude into Your Personal Real Estate AI Assistant
FollowUpAce's MCP integration exposes your Follow Up Boss CRM to Claude (via https://followupace.com/mcp) and ChatGPT (via SSE at https://followupace.com/api/mcp/sse/) through a single URL. Pro-plan users get 215 tools covering pipeline analysis, lead nurture, Zillow speed-to-lead checks, Fair Housing compliance scanning, and more — all queryable in plain English without leaving your AI assistant.
What is the Model Context Protocol (MCP) and why does it matter for real estate agents?
MCP — the Model Context Protocol — is an open standard that lets AI assistants like Claude and ChatGPT call external tools and read live data, rather than relying solely on their training knowledge. When an MCP server sits in front of your CRM, the AI can look up real contact records, check pipeline stages, and take actions on your behalf, all from a natural-language prompt.
For real estate agents, this matters because the most time-consuming CRM work is not data entry — it is synthesis. Who needs a follow-up today? Which deals have sat in the same stage too long? Which Zillow leads came in while you were at a showing? Answering those questions normally means clicking through multiple screens. With an MCP-connected assistant, you ask once and get a prioritized answer drawn from your actual data.
Before MCP, connecting an AI to a CRM required custom API integrations, developer time, and ongoing maintenance. MCP standardizes the handshake so any MCP-aware AI client can connect to any MCP server via a single URL — no custom plumbing required.
How does FollowUpAce's MCP integration work?
FollowUpAce runs an MCP server that sits between your AI assistant and your Follow Up Boss account. Once you add the connector URL in your AI client, every conversation has read and write access to your live CRM data through a catalog of 215 tools (as documented in mcp-server/src/index.ts, line 206).
There are two endpoints depending on which AI client you use:
| AI Client | Protocol | Connector URL |
|---|---|---|
| Claude (Anthropic) | Streamable HTTP | https://followupace.com/mcp |
| ChatGPT | Server-Sent Events (SSE) | https://followupace.com/api/mcp/sse/ |
Access is gated at the Pro plan level (embed.js, line 4311: proRequired: planType !== 'pro'), which sits at $55 per seat per month when purchased individually, or can be provisioned by an admin for the team.
How do I connect Claude to my Follow Up Boss CRM through FollowUpAce?
The setup takes under five minutes. Here is the step-by-step process:
- Log into your FollowUpAce dashboard and confirm your account is on a Pro plan.
- Navigate to the MCP Integrations section. You will find both connector URLs displayed there, ready to copy.
- Open Claude.ai (or your preferred Claude client that supports MCP). Go to Settings › Integrations and paste the FollowUpAce connector URL:
https://followupace.com/mcp. - Authorize the connection. FollowUpAce authenticates via your existing Follow Up Boss credentials — no separate API key needed.
- Start a new conversation and ask something like: "Show me all leads that came in from Zillow in the last 48 hours and flag any that have not received a response yet."
For ChatGPT, the process is the same but you paste the SSE endpoint (https://followupace.com/api/mcp/sse/) into the ChatGPT connector settings instead.
What can Claude actually do inside Follow Up Boss through FollowUpAce?
The 215-tool catalog spans the full lifecycle of a real estate transaction. Below are the categories most agents reach for first:
Pipeline health and deal management
The pipeline-health-check tool (verified at mcp-server/src/index.ts, line 4198) analyzes your deal pipeline end-to-end: stuck deals, overdue stage transitions, estimated revenue forecast, win/loss rate, and deals needing immediate attention. You can specify a days_stuck_threshold — for example, flag any deal sitting in the same stage for more than 21 days — and Claude will return a prioritized list with recommended next steps.
Ask Claude: "Run a pipeline health check and tell me which deals are most at risk of falling through this month."
Lead nurture and re-engagement
The lead-nurture-optimizer tool (verified at mcp-server/src/index.ts, line 4230) identifies contacts who have gone cold and surfaces personalized re-engagement action plans. You can filter by stage (New, Prospect, Active Buyer, Past Client) and set a minimum days-since-contact threshold. The result is a ranked list of who to call today and why, drawn from your actual CRM data.
This connects naturally with the agentic suggestion layer in FollowUpAce, which can push prioritized follow-up nudges directly into the FUB interface.
Zillow speed-to-lead monitoring
Speed-to-lead response time is one of the most researched conversion drivers in residential real estate. FollowUpAce exposes a dedicated lead_speed_to_lead_check tool (verified at mcp-server/src/zillowTools.ts, line 377) that checks whether a specific contact is still within the critical response window, returns the lead age in minutes, a priority classification based on timing signals, and the recommended immediate action. For team-wide analytics — average first-response time and percentage of leads reached within five minutes — use the companion lead_analytics_summary tool. You can run the per-contact check on incoming leads to decide response urgency.
See the full Zillow speed-to-lead playbook for context on how to set up your response benchmarks.
Fair Housing compliance scanning
Every message drafted inside Claude can be scanned against Fair Housing Act rules before it is sent. The scan_message_for_compliance tool calls scanForComplianceViolations() in chat-app/utils/complianceGuard.js (line 293), which checks text against prohibited terms under 42 U.S.C. 3601-3619 — covering race, color, religion, national origin, sex, familial status, and disability — plus state and local extensions and licensing boundary rules.
Rather than remembering every prohibited phrase yourself, you can paste a draft message into Claude and ask: "Check this email for Fair Housing compliance issues." Learn more about how FollowUpAce handles this on the compliance page.
Team performance and coaching
The team-performance-snapshot tool reports lead response times, deals per agent, task completion rates, pipeline value by agent, and coaching opportunities — filterable by time period (last 7 days through year-to-date) or by individual agent. Team leads can ask Claude to compare agent performance across quarters without pulling a manual report.
How does the Ace Score and Ace Trove layer connect to the MCP integration?
The MCP integration sits on top of FollowUpAce's broader data layer. As Ace Trove analyzes contacts, it writes structured signals into Follow Up Boss as custom fields. These fields are then readable through the MCP tools, so Claude can reference current Ace Score, Ace Tier, and AI-derived fields when formulating recommendations.
The current field system (verified in chat-app/shared/aceIntelligenceFieldsConfig.js) includes:
The pre-consolidation free fields (retired June 2026 — see the update note above):
- Ace Score — engagement score 0–100
- Ace Tier — Hot / Warm / Cool / Cold / Dormant
- Ace Status — New / Contacted / Engaged / Active Client / Closed / Nurturing / Dormant
- Ace Response Time — minutes to first response after lead creation
- Ace Velocity Score — engagement velocity derived from 7-day and 30-day event windows
- Ace Days Since Inbound — days since the contact last reached out
- Ace Preferred Channel — SMS / email / call, recency-weighted
The pre-consolidation paid Ace Trove fields (also consolidated — today: Expected GCI est., Opportunity, Seller Tier, Last Analyzed):
- Ace Lead Type — Buyer / Seller / Investor / Buyer-Seller / Unknown
- Ace Property Profile — AI-synthesized preferences from IDX search data
- Ace Search Area — geographic search areas from IDX activity
- Ace Buyer Readiness — Ready Now / Actively Looking / Researching / Passive / Unknown
- Ace Lead Summary — AI-generated summary from notes and activity patterns
- Ace Next Action — AI-recommended next step for this contact
- Ace Seller Score — propensity-to-sell score 0–100 from property and behavioral signals
- Ace Seller Tier — Very High / High / Moderate / Low / None
When Claude asks the MCP server about a contact, these fields are part of the context it receives, giving the AI a richer picture than raw CRM notes alone. For more on what Ace Trove analyzes, see the Ace Trove overview.
What does FollowUpAce MCP cost, and what plan do I need?
MCP access requires a Pro seat. Per-seat pricing is:
| Plan | Per Seat / Month | MCP Access | Voice Chat |
|---|---|---|---|
| Free | $0 | No | No |
| Regular | $25 | No | No |
| Pro | $55 | Yes | Yes |
Team leads who want to run Ace Trove analysis across their entire contact database can add an account-level Ace Trove subscription on top of per-seat pricing. The Ace Trove tiers start at $49/month for up to 5,000 contacts (Starter) and scale to $899/month for up to 500,000 contacts (Enterprise), with Growth ($99), Team ($199), Brokerage ($349), and Scale ($549) tiers in between — as verified in chat-app/config/accountAiBillingConfig.js, lines 64–69.
How is talking to Claude through MCP different from using FollowUpAce's built-in chat?
The built-in FollowUpAce chat assistant runs inside your Follow Up Boss interface and is tuned for quick, contextual answers about a single contact or conversation. The MCP integration is for a different use case: open-ended, multi-step analysis across your entire database from inside the AI client you already use day-to-day.
| Capability | Built-in FUB Chat | Claude via MCP |
|---|---|---|
| Scope | Single contact / conversation context | Entire database + pipeline |
| Where you work | Inside Follow Up Boss | Claude.ai or ChatGPT |
| Pipeline health check | No | Yes |
| Team performance report | No | Yes |
| Fair Housing compliance scan | Yes (in-context) | Yes (on-demand) |
| Combine with other AI prompts | No | Yes — mix CRM queries with any other task |
In practice, most Pro agents use both: the in-app chat for quick contact-level questions during a showing, and Claude for morning planning sessions where they want a full account overview.
What prompts work well with FollowUpAce MCP and Claude?
The most productive prompts are specific and outcome-oriented. Here are patterns that work well with the available tools:
Pipeline review: "Run a pipeline health check and flag any deal that has been in the same stage for more than 21 days. For each one, suggest a realistic next step."
Cold lead reactivation: "Find all contacts I have not reached out to in the last 45 days who are tagged as Active Buyer, and draft a short re-engagement text for each one."
Zillow lead triage: "Check all Zillow leads from the past 72 hours and tell me which ones I have not responded to yet, sorted by how long they have been waiting."
Compliance check: "Before I send this email, check it for any Fair Housing compliance issues." [paste draft]
Team coaching: "Show me a team performance snapshot for the last 30 days. Which agent has the slowest average lead response time, and what deals do they currently have in the pipeline?"
Morning briefing: "Give me my top five contacts to prioritize today, based on Ace Score, days since last contact, and any overdue tasks."
Is FollowUpAce the only MCP connector for Follow Up Boss?
MCP is an open standard, so in principle any developer can build a connector for any system. What differentiates FollowUpAce's connector is that it is purpose-built for the real estate workflow on top of Follow Up Boss — with tools like speed-to-lead checks, Fair Housing compliance scanning, Seller Score analysis, and the Ace Trove field layer already wired in. A generic FUB API wrapper would expose raw endpoints; the FollowUpAce MCP server exposes curated, task-specific tools that return actionable answers rather than raw records.
For teams evaluating alternatives, the competitor comparison pages include an overview of which platforms offer AI-driven CRM automation and where the capability gaps lie.
Frequently asked questions about FollowUpAce MCP
Does the MCP integration write back to Follow Up Boss, or is it read-only?
The connector supports both reads and writes. You can ask Claude to update a contact's stage, log a note, create a task, or send a message — all of which flow through to your live FUB account.
Is my CRM data sent to Anthropic or OpenAI?
When Claude or ChatGPT calls an MCP tool, the tool request goes to the FollowUpAce server, which fetches data from Follow Up Boss and returns it in the same conversation session. The AI model receives only what the tool returns — it does not have persistent access to your database outside of an active conversation. Review Anthropic's and OpenAI's data handling policies for details on how conversation data is treated by each provider.
Do I need a developer to set this up?
No. The connector is a URL you paste into your AI client's settings. FollowUpAce handles authentication through your existing Follow Up Boss login.
Can I use this with Claude on mobile?
MCP support in mobile AI clients varies by app and version. The desktop web versions of Claude.ai and ChatGPT have the most consistent MCP support as of mid-2026.
What happens if my Follow Up Boss API connection expires?
The MCP server will return an authentication error to Claude, which will surface it in the conversation. You re-authorize through the FollowUpAce dashboard — the same flow as refreshing any OAuth connection.
Background: AI real estate assistant — capabilities and setup.