- Brand
- Follow Up Boss
- Category
- Sales & CRM
- Primary Subcategory
- Real Estate Agent & Broker Software
Integration details
Description
Follow Up Boss helps real estate agents and teams work with their CRM in chat. Connect your existing Follow Up Boss account to review contacts and recent notes, plan follow-up from tasks, appointments and smart lists, draft personalized outreach, and analyze agent activity and conversation counts. Prepare supported CRM changes for review and confirm them in the review interface before saving. Email and SMS messages remain drafts; the plugin does not send messages or place calls. Requires Follow Up Boss sign-in and applicable account permissions. Results reflect the records and coverage available to your connected account.
- Integration type
- Plugin
- Verification status
- Not applicable
- Platform
- ChatGPT
- Primary Subcategory
- Real Estate Agent & Broker Software
- Secondary Subcategories
- None listed
- Brand
- Follow Up Boss
- Access
- Account required
- First tracked
- 2026-11-20
- Tool count
- 89
- Market
- US
The Primary Subcategory used for this profile’s headline score.
Other Subcategories where the Integration is listed.
Not discoverable in ChatGPT yet
ChatGPT now suggests Plugins on its own when they match a user's request, but not every Plugin is being surfaced yet.We haven’t seen Follow Up Boss suggested yet. Sign up and we’ll track it, and your Discoverability Score will appear here once ChatGPT starts surfacing it.
What discovery looks like

Get alerts for Follow Up Boss
Get updates when Follow Up Boss’s Discoverability Score or category rank changes.
Competing in ChatGPT Real Estate Agent & Broker Software
View Category89 tools agents can invoke
Read the progress of one bulk contact update the widget started. Requires the peopleJobId and the private statusToken the widget received with it; a job without a valid token from this user is refused. Makes one Follow Up Boss job status read and no writes.
mcp_app_get_bulk_people_job
Pause the one enrollment the user picked in a reviewed Follow Up Boss action-plan pause, after the user confirms it in the pause widget. One enrollment per confirmation; ask again for the next one. Requires proposalId, its private confirmationToken, confirmed=true and selectedEnrollmentIds holding exactly one enrollment from that review. Pauses only that enrollment, once, and posts the reviewed reason note once. Never resend an uncertain pause; check Follow Up Boss instead.
mcp_app_apply_pause_action_plan
Save one reviewed Follow Up Boss appointment change after the user confirms it in the appointment widget. Requires proposalId, its private confirmationToken and confirmed=true. Makes exactly one write and never sends invitations. Never resend an uncertain write; check Follow Up Boss instead.
mcp_app_apply_appointment_update
Log one reviewed call on a Follow Up Boss contact after the user confirms it in the call log widget. Requires proposalId, its private confirmationToken and confirmed=true. Saves exactly the reviewed call once; it never places a call. Never resend an uncertain log; check Follow Up Boss instead.
mcp_app_apply_log_call
Save one reviewed Follow Up Boss interaction update after the user confirms it in the update widget. Requires proposalId, its private confirmationToken and confirmed=true. Makes exactly the reviewed writes once. Never resend an uncertain write; check Follow Up Boss instead.
mcp_app_apply_interaction_update
Apply one immutable Follow Up Boss deal proposal after the user explicitly confirms in the deal review widget. Requires proposalId, its private confirmationToken and confirmed=true. Performs one native create or field-scoped update with readback, guarded by auth, feature permission and MUTATIONS_ENABLED. Returns a receipt without rendering another widget. Never resend an uncertain write or substitute new fields. Inspect the native deal after any lost response; simultaneous confirmations are not deduplicated.
mcp_app_apply_deal_update
Delete one reviewed Follow Up Boss record after the user confirms it in the delete widget. Requires proposalId, its private confirmationToken and confirmed=true. Deletes a note, task or deal. Contact deletion is unsupported here, including older contact confirmation tokens; open the contact in Follow Up Boss instead. It re-reads the record first and deletes nothing if it changed since the review; a replay finds it already deleted and writes nothing. Never resend an uncertain delete; check Follow Up Boss instead.
mcp_app_apply_delete_record
Pause or resume the signed-in Follow Up Boss agent's leads on Follow Up Boss and Zillow, after the user confirms it in the lead pause card. Requires proposalId, its private confirmationToken, confirmed=true and action; a pause also needs expiresAt, in the future and at most 48 hours out. A new pause replaces the current one. Returns the status read back from Follow Up Boss. Never resend an uncertain change; check the status instead.
mcp_app_apply_lead_pause
Apply a reviewed bulk contact update only after the user explicitly confirms it in the widget. Requires proposalId, the private confirmationToken, confirmed=true and personIds: the reviewed contacts the user left ticked. Auth, account permissions and mutation switches are rechecked; contacts outside the review are refused. The ticked contacts are re-read as the user: if any is no longer visible the apply is refused, contacts that already have the change are left out, and when none is left nothing is sent (nothing_to_change). Otherwise sends exactly one Follow Up Boss mass update and returns its job ID and progress. Tokens never enter the transcript. Never retry or replay; an uncertain outcome is checked in Follow Up Boss.
mcp_app_apply_bulk_people_update
Apply the immutable contact assignment proposal only after the user explicitly confirms its rendered review. One contact per confirmation; ask again for the next one. Requires proposalId, private confirmationToken and confirmed=true. Auth, CSD permission, mutation switches, active target roles, contact/pond access and stale routing are revalidated. Only the frozen reviewed contact can change; returns its verified, failed or uncertain outcome without a second widget. Tokens never enter the transcript. Completed receipts expose no confirmation token. Signed confirmations do not guarantee exactly-once dispatch; submit only once. Does not reassign existing tasks, send customer messages or start action plans. Never retry or replay uncertain outcomes.
mcp_app_apply_contact_assignment
Create the one reviewed Follow Up Boss contact the user confirmed in the contact widget. One contact per confirmation; ask again for the next one. Requires proposalId, its private confirmationToken, confirmed=true and selectedRowIds holding exactly that review's one row. The row is created at most once, exactly as reviewed; if it now matches an existing contact, or its assignee or referring contact is no longer valid, nothing is created and the user reviews again. A partial status means the create failed, could not be confirmed, or was saved without its referring link. Never resend an uncertain create; check Follow Up Boss instead.
mcp_app_apply_create_contact
Apply the frozen Follow Up Boss task review only after explicit user confirmation in its widget. One task per confirmation; ask again for the next one. Requires proposalId, private confirmationToken and confirmed=true; no other task ID or new field value can be supplied here. Rechecks authentication, feature permission, mutation switches, native access and the reviewed fields, then patches only the requested completion, date or assignee fields and reads the task back. Returns its verified, failed or uncertain result in the same widget. Concurrent confirmations and retries of the same live proposal reuse its result. After an uncertain outcome inspect FUB before preparing another review. Does not create or delete tasks, send customer messages or start action plans.
mcp_app_apply_task_update
Draft starting a Follow Up Boss action plan for one exact contact as a card, for requests to start, apply, add or enroll a contact in an action plan. This app never starts action plans and this tool writes nothing: the public API has no draft enrollment, so the card names the contact and the plan (with its step count when Follow Up Boss returns steps), shows whether the contact is already enrolled and with what status, and links to the contact in Follow Up Boss, where the user starts the plan. Supply personId and one of actionPlanId or a nonblank planName (one is required); a name matching several plans returns them to choose from, and an unknown contact or plan says so. To stop a running plan, use pause_action_plan.
mcp_app_render_action_plan_start
Prepare an editable meeting-note draft for one exact Follow Up Boss contact. This is read-only and makes no FUB API call. The widget requires explicit review before it posts the note: it prepares an official interaction_update review of exactly that note and saves it once with apply_interaction_update; posting is unavailable when the interaction-update CSD feature is missing or MUTATIONS_ENABLED is false.
mcp_app_draft_meeting_note
Draft a Follow Up Boss SMS or email for the user to review. This tool never sends and never writes; it only returns the draft. In the widget the user can copy the draft; for an email, the user can also save it as a draft on the contact in Follow Up Boss after confirming, which the separate apply_save_email_draft tool does (an SMS has no Follow Up Boss draft, so it is only copied). Where the user's account is enabled to send, the widget can also send exactly this draft after the user confirms, through the separate app-only send_sms or send_email tool; otherwise sendNote says why (sendBlock is set when sending is off for the whole account rather than for this draft), and the user sends it from Follow Up Boss. Never call send_sms or send_email yourself. If the result status is login_required, complete Follow Up Boss sign-in first.
mcp_app_draft_message
Draft personalized, review-only outreach messages for a bounded set of authoritative Follow Up Boss person IDs. The tool fetches canonical people data plus bounded recent text-message history and notes, then renders one draft per resolved person. Person IDs may come from any independently gated FUB cohort capability; this tool does not call or depend on a cohort tool. Nothing is sent or written. The Follow Up Boss public API cannot send messages, so the user sends any chosen draft from Follow Up Boss.
mcp_app_draft_outreach_messages
Purpose: render the exact selected FUB follow-up lead with an editable manager-authored timeline-note suggestion. Source: one canonical selectedLead copied unchanged from render_agent_follow_up_leads plus a visibly separate draft and annotations. Category: render-only; this makes zero FUB API calls. The Post action is disabled unless the account has fub-mcp-interaction-update, MUTATIONS_ENABLED is true, and the user explicitly approves; posting prepares an official interaction_update review of exactly that note and saves it once with apply_interaction_update. It never uses a second note mutation, sends SMS/email, notifies or contacts the assigned agent, or posts during drafting.
mcp_app_draft_manager_follow_up_note
Find a Follow Up Boss contact by name, email, or phone and render a lead card with disambiguation when multiple contacts match. Pass focus when the user asked about one aspect only: methods for the other phone numbers and emails, tags, fields for custom fields, relationships, or missing for fields that are Not set. A name that matches no contact is also looked up among related contacts (spouse, family, co-buyer and the like) and returns the primary contact with matchedVia naming that relationship; notes, calls or updates about the related person belong on that primary contact. The result states the match outcome (status, matchCount) and each candidate's id and name; when several contacts match, ask the user which one. If the result status is login_required, the caller is not authenticated; complete Follow Up Boss sign-in, then retry this tool.
mcp_app_find_contact
Read authoritative account vocabulary for stages, timeframes, users, custom fields, pipelines, and email/text template identity through fully paginated FUB API requests. The tags resource is always unavailable: the public API has no account tag list, so read tags from each contact. Optionally select fixed resource groups; omission reads all. Template content is never returned. Each endpoint reports available, empty, forbidden, or unavailable status and explicit completeness; partial data remains marked incomplete. The tool renders no UI, caches no account data, and performs no writes. The account must be authenticated and have database-read access.
mcp_app_get_account_schema
Count calls and occurred appointments per visible Follow Up Boss agent from the public API (GET /users, /calls and /appointments) and render one card per agent. Sent texts are always unavailable (texts is null): the public API has no account-wide text list, so never invent a text count. The default range is the last seven calendar days in the signed-in user's FUB time zone. Supply only an authoritative FUB user ID in agentId; when it is unclear which agent the user means, ask in plain text before calling. Scans are page-capped; when countsComplete is false every count is a lower bound. This is read-only. Do not describe appointments as appointments created by the agent.
mcp_app_get_agent_activity_counts
Count an agent's qualifying Follow Up Boss calls in an inclusive UTC period. There is no widget: answer in chat with the count and the period. A conversation is an active public call lasting at least 120 seconds, attributed by userId. The requested agent ID is validated against public /v1/users, and calls are scanned through public /v1/calls only. A partial scan is a lower bound and must never be presented as an exact total. This is read-only.
mcp_app_get_agent_conversation_count
List Follow Up Boss leads the signed-in user can claim: open First to Claim offers made to them, reassignment offers included, and leads in up to 5 ponds they are a member of, with each lead's source, created date, and price and area when FUB gives them before a claim, plus a FUB link for pond leads. Ponds they see only through team members are not included, and the text says how many other ponds exist. pondId limits it to one pond (pond IDs come from get_teams with resource ponds), reads that pond even when the user is not a member (FUB still limits what they can see) and skips First to Claim offers. createdAfter (YYYY-MM-DD) limits it to recent leads. Use it for questions like what leads can I claim, what's in the Nurture pond, or are there unclaimed leads from today. It never claims, assigns or moves leads: claiming happens in Follow Up Boss, and assigning a lead or moving it into a pond is render_assignment_review. First to Claim offers fall back after about 15 minutes, so none being open is normal. The account must be authenticated. It renders no UI and performs no writes.
mcp_app_get_claimable_leads
Look up the phone number to call one Follow Up Boss contact. This tool never starts a call: the Follow Up Boss public API has no dial or call-bridge endpoint, so it returns the contact's phone, a tel: link and the View in FUB link for the user to call from Follow Up Boss. Read-only. If the result status is login_required, complete Follow Up Boss sign-in, then retry.
mcp_app_get_contact_phone
Answer "who should I call today?" with a bounded, read-only Follow Up Boss queue using the deterministic stage-aware-no-next-action-v2 rule and the fub-mcp-needs-follow-up capability. Results are the signed-in user's own assigned leads by default, which is what "who should I call today?" asks for; no extra call is needed to scope it. Pass scope="team" for the whole account, or assignedUserId / assignedTo for one named agent. The rule combines stage-aware cadence, local-today open tasks, recent engagement, action plans, the top-funnel per-lead contact-attempt cap, and unanswered inbound calls/texts. dueWindow=today includes today and overdue leads; unanswered inbound leads rank first. Incomplete source coverage is exposed as partial; this tool never writes or contacts anyone and a bounded result is not an exhaustive account total.
mcp_app_get_leads_needing_followup
Analyze a bounded set of exposed Follow Up Boss inbound call, text, and email activity and report whether a later exposed outbound response exists. Defaults to inbound activity with no later exposed response in the last 14 days. A response is only claimed when an outbound timestamp exposed by FUB is strictly later than the latest inbound timestamp; missing fields are not proof that no outside or unobservable response exists. Use responseStatus=responded or any to inspect other statuses. This is distinct from unworked-new-lead speed-to-lead checks and stage/age/cadence follow-up queues. Repeat with coverage.nextOffset for bounded continuation or reuse the same filters to refresh. Read-only; partial coverage must not be presented as an exact account total.
mcp_app_get_messages_by_response_status
Return per-agent counts projected from the same deterministic stage-aware-no-next-action-v2 collector as get_leads_needing_followup. This read-only count view combines stage, age, engagement, open-task, action-plan, tag, and follow-up-cap rules without inventing a native FUB aggregate. Coverage is bounded and explicitly partial when any source is incomplete. Use get_leads_needing_followup for the underlying lead list; do not use this for unworked-new-lead or message-response-status questions.
mcp_app_get_needs_follow_up_counts
Use this when the user wants to work their Follow Up Boss database, find who to follow up with today, review stale leads, triage past clients, or identify outreach opportunities across multiple contacts. This read-only tool returns scored candidates with concise evidence and suggested next touches. Do not use for single-contact lookup; use find_contact instead. It neither drafts nor sends messages. After the user chooses a candidate and message, draft it with draft_message, which only drafts: the user reviews the draft and sends it from Follow Up Boss. If Zillow tools are available, compose with the returned address and inquiryProperty fields: address is the contact's home for an equity-check angle, and inquiryProperty is a home they wanted for relist, price-drop, or similar-property angles.
mcp_app_get_outreach_opportunities
Find a bounded set of stale Follow Up Boss people and enrich each with up to three recent notes plus deterministic reengagement signals. This authenticated, read-only data tool uses oldest activity first and reports partial note coverage; it does not rank beyond that FUB sort, render UI, write notes, or contact anyone.
mcp_app_get_reengagement_candidates
Retrieve a bounded page of people in a Follow Up Boss smart list. Use a smart-list ID when known; otherwise omit the identifier to discover lists or pass a name for exact-then-substring selection. Reuse pagination.next with the same list and page size to continue native ordering. Column configuration is returned only when exposed by FUB; contact rows contain summary fields. The widget handles list selection, pagination, and refresh. Use render_smart_list_people for an exact curated subset. This MCP cannot tag or update a whole smart list. When bulk_update_people is available, it can make one bounded change (add or remove tags, change stage, reassign, or assign a pond) to up to 100 contacts by exact ID from this list, which the user reviews and confirms in its widget. For more contacts, other changes, or when bulk_update_people is not available, say up front that it isn't supported here, then offer to tag contacts one at a time with interaction_update or to use Follow Up Boss's own smart-list bulk actions. Do not show this list for the user to select from for a bulk action; the bulk_update_people review lets the user untick contacts. For a bulk change, link the Follow Up Boss People list so the user can do it there: https://app.followupboss.com/2/people
mcp_app_get_smart_list_people
Compare Follow Up Boss lead sources using a bounded, read-only activity-window scan of people and active/archived deals. Returns people, contacted, stage, deal, closed-deal, value, contact-rate, and dealsPerScannedPerson metrics sorted by dealsPerScannedPerson. dealsPerScannedPerson is NOT a conversion rate: unwindowed first-page deals / activity-window people; it may exceed 1.0. Optional exact source names and assignedTo narrow the report; incomplete scans are explicitly lower bounds. The account must be authenticated and have the fub-mcp-source-performance capability. It renders no UI and performs no writes.
mcp_app_get_source_performance
Read the account's own stage and timeframe vocabulary in one tool invocation using three bounded FUB API reads: people stages from GET /stages, deal stages from GET /stages?type=deals, and timeframes from GET /timeframes, each ordered by orderWeight where FUB supplies it. Take no input. Call it before judging whether a person's stage or timeframe is out of step, so the comparison uses this account's names and ordering instead of assumed ones. It renders the account's contact stages, in that order, as a card; deal stages and timeframes stay in the reply. It changes no stage or pipeline configuration, and reports each read's status (available, empty, forbidden, or unavailable) and warning in coverage.endpoints rather than failing when one read is denied. The account must be authenticated.
mcp_app_get_stage_definitions
Count qualifying Follow Up Boss calls for every supplied team member in an inclusive UTC period, including agents with zero conversations. There is no widget: answer in chat with each agent's count. A conversation is an active public call lasting at least 120 seconds, attributed by userId. Supply authoritative agentId and agentName pairs without name matching or substitution. Calls are scanned through public /v1/calls only. A partial scan contains lower bounds and must never be presented as exact totals. This is read-only.
mcp_app_get_team_conversation_counts
Purpose: retrieve a bounded manager-visible Follow Up Boss daily brief with active-lead, yesterday-conversation, and deterministic fub-mcp-team-daily-brief needs-follow-up coverage per exact agent ID. Source: public FUB users/teams/people/calls plus the shared stage-aware-no-next-action-v2 collector; Category: data-only read with no widget; repeat to refresh. Supply teamLeadUserId only when it is authoritative; after review call render_team_daily_brief with only returned records. Partial scans remain partial and unavailable totals are withheld; never substitute unworked-new-lead or unanswered-inbound metrics.
mcp_app_get_team_daily_brief
Report how a Follow Up Boss team is performing over a date range: per agent leads assigned, new leads, speed to action (minutes from lead creation to the agent's first action on the lead, the same as the Agent Activity report: a call, text, email, note or Zillow message, a task, or a lead edit such as a stage change, counted only for automatically created leads) and leads not acted on, each with the previous period of the same length alongside, and a weekly trend of new leads, speed to action and not acted on for each agent and the team (trendWeeks, default 4 weeks ending at endDate; the last week can still change). Speed to action comes with the count of leads that have a value, and for the current period also the median and slowest when Follow Up Boss returns them. It also returns a current snapshot of each agent's leads by lead stage, with how many have sat in their stage longer than stuckDays. The snapshot is not limited to the date range and has no previous period. Use it for questions like how is my team doing, who is slowest to respond to new leads (rank by median when present), or where are my leads getting stuck. Results follow Follow Up Boss report permissions for the signed-in user: admins see every agent, team leads see their team, agents see only themselves. Activity numbers match the Follow Up Boss Agent Activity report with its default Web leads filter, and the coverage notes list the exceptions, such as leads still assigned to deleted users or lenders. The team average speed to action (meanMinutes) is the average of the agents' averages, as in that report, and the team median and slowest are over all the team's leads. The account must be authenticated. It renders no UI and performs no writes.
mcp_app_get_team_performance
Purpose: answer who a selected agent should call today with an urgency-ordered lead follow-up queue. Source: the authoritative exact FUB user ID plus bounded people, open-task, action-plan, recent-engagement, call, text, and capped tag-detail reads using the shared stage-aware-no-next-action-v2 rule. dueWindow=today includes today and overdue leads; unanswered inbound calls/texts rank first. Category: data-only read with no widget; repeat only with the returned opaque nextCursor, same agentId, dueWindow, and explicit limit. The exact agent ID must come from the team brief handoff; never infer it from a name. After exploration call render_agent_follow_up_leads exactly once. It never writes or contacts anyone.
mcp_app_get_agent_follow_up_leads
List Zillow leads at risk because they have had no two-way conversation in the last 15 days: Follow Up Boss leads received from Zillow Preferred, Zillow Flex or Zillow in the last 90 days, where received is the latest Zillow inquiry (a re-inquiry on an older contact counts), with neither a connected call (2 minutes or more, either direction) nor both an agent message and a lead message in the last 15 days, most urgent first. Calls, texts, emails, Zillow messages and inbox messages count by direction, in either order, and unanswered calls count by direction too. Only the lead's own messages count as the lead's, and appointment reminder texts count as the agent's. Automated messages from Follow Up Boss count on neither side, but a lead's out-of-office auto-reply counts as a reply. A lead is not at risk until 15 days after it was received. Leads in the Trash stage are not included. Each lead has its Zillow program, date received, latest agent message and latest lead message (channel and date), latest connected call (direction and date), assigned agent and a FUB link. Use it for questions like which Zillow leads are at risk, or any Preferred leads I haven't talked to in two weeks. Results follow the signed-in user's FUB visibility, so team leads see their agents' leads, and assignedUserId narrows to one agent. It is not for new leads that were never worked (get_unworked_new_leads) or the general follow-up queue (get_leads_needing_followup). Texts sent from a personal phone are not recorded, and the scan is bounded, so the count is a minimum when coverage is incomplete. It needs FUB's inquiry filter and conversation signals, and returns an error without them. The account must be authenticated. It renders no UI and performs no writes.
mcp_app_get_at_risk_zillow_leads
Count the signed-in user's overdue, incomplete Follow Up Boss tasks and return a link that opens their overdue tasks in Follow Up Boss with the "Clear Overdue Tasks" confirmation already open. Use it when the user asks to clear, delete, remove or complete all or many overdue tasks. This MCP takes no bulk task actions: do not attempt bulk completion one task at a time through render_task_review instead. Confirming in Follow Up Boss deletes (does not complete) every overdue task assigned to the user and cannot be undone; Follow Up Boss has no bulk "mark complete". Takes no input, makes no writes and renders no UI. When nothing is overdue it returns no link. The account must be authenticated.
mcp_app_get_clear_overdue_tasks_link
Return the authenticated Follow Up Boss user by calling /me with the active FUB credential.
mcp_app_get-current-user
Identify Follow Up Boss leads with deterministic declining-engagement evidence. Showing Homes and Submitting Offers leads qualify after 14+ days without recorded activity and without a recent IDX visit; Met with Customer leads qualify after inactiveDays (default 30) without recorded activity. Optional inactiveDays (1-365), stage, source, and assignedTo are passed through to the FUB people scan. The bounded, read-only scan hits the stale lastActivity slice first, includes recorded response-rate evidence, canonical FUB links, and explicit partial-coverage details. There is no widget: answer in chat with each lead's name, evidence and FUB link, and say when coverage is partial. Use assignedTo to keep the review to one agent.
mcp_app_get_declining_engagement_leads
Brief the signed-in Follow Up Boss user on their own day in one call: today's FUB appointments (synced Google or Outlook calendar events are not included), tasks due today, overdue tasks, leads with unanswered texts or emails from the last 14 days (calls are not counted) and the top leads flagged for follow-up, each with its full count and its first 5 items, with FUB links on contact-linked tasks, messages and leads, dated in the user's FUB time zone. Set remainingOnly for what's left today, which drops completed tasks and appointments that have ended. Use it for questions like what does my day look like, give me my morning brief, or what's left for today. It covers only the signed-in user's own day: for a manager's brief on the team use get_team_daily_brief, for tasks only, such as what's on my plate today, use get_tasks, and for the schedule only or another day, such as my schedule tomorrow, use get_appointments. A section whose read fails, is incomplete or takes over 20 seconds is marked partial or unavailable, while the other sections still return. The account must be authenticated. It renders no UI and performs no writes.
mcp_app_get_my_daily_brief
Brief on recently created Follow Up Boss leads with no exposed agent outreach. 'Worked' means an outgoing agent call, text, or email exposed by FUB; lead-driven activity such as site visits or incoming messages does not count. Use this for speed-to-lead briefing on new leads, not the broader needs-follow-up cadence or unanswered-inbound analysis. The scan is read-only and bounded. There is no widget: answer in chat with each lead's name, assigned agent and canonical FUB link, and say when coverage is partial; repeat with the same lookbackDays for a fresh check.
mcp_app_get_unworked_new_leads
Identify which Follow Up Boss contact a recording or note is about from up to four mentioned names, with optional appointment-overlap evidence. A name that matches no contact is also looked up among related contacts (spouse, family, co-buyer and the like) and returns the primary contact with matchedVia naming that relationship; notes, calls or updates about the related person belong on that primary contact. This read-only tool returns bounded, ranked canonical candidates and renders a confirmation picker. Use it when a recording or transcript needs a contact identity; wait for the widget or a follow-up call to return confirmed.personId before continuing to any contact-specific workflow. A candidates list is not confirmation, and this tool never silently selects a person. It may be repeated only to confirm one exact candidate or choose None of these. Appointment evidence is best-effort and any unavailable coverage is labeled in limitations.
mcp_app_identify_contact
Inspect Follow Up Boss smart-list, automation, and action-plan diagnostics without writing. Use this for questions like why a lead is in a smart list, which automations or action plans are active for a contact, or whether automation diagnostics are permission-limited. This stage is read-only; admin writes such as applying, pausing, or updating automations are intentionally not performed.
mcp_app_inspect_fub_diagnostics
Read a bounded set of recent notes from one exact Follow Up Boss contact. Use a personId from find_contact or another authoritative FUB result. This is read-only and never creates or edits notes. There is no widget: answer in chat with each note's author, date and text, newest first, and say when the reply is narrowed to one author or a date range.
mcp_app_list_notes
Open the Follow Up Boss workspace, a view of a smart list's contacts. It opens in the side panel where the host supports one, a full-height view that stays open beside the chat, and inline in the chat otherwise. Pass list with a smart-list name, matched exactly and then as a substring like get_smart_list_people's smartListName, to open on that list; omit it to show the user's smart lists to choose from. Pass personId with a contact's FUB person ID to point the workspace at that contact. Use it when the user wants to open their workspace or work through a list in the workspace, such as open my workspace or open my New Leads list in the workspace. For a one-off answer about a list in chat, use get_smart_list_people. This tool only opens the view and makes zero FUB API calls; the view reads the list with get_smart_list_people, and it writes only when the user clicks Save to Follow Up Boss on a staged email, which saves it as a Follow Up Boss email draft and sends nothing. The account must be authenticated.
mcp_app_open_workspace
Read authorized Follow Up Boss contacts without opening a widget. Supply personId alone from a resolved contact to read complete canonical contactDetails: all contact methods, profile, relationships, assignments and typed custom values. Search by name or other filters first when the ID is unknown; ambiguous names require user selection, never guess an ID. Missing fields are not proof of blank values or permission denial; stage-change provenance and channel preference are unavailable unless explicitly source-backed. Schema coverage may be partial. This data-only tool renders no UI. Without personId, retrieve bounded canonical people summaries by tags, broad query, specific person fields, or date windows. createdAfter/createdBefore filter record creation; updatedAfter/updatedBefore filter record modification. Use exclusive ISO timestamps with Z or an explicit UTC offset; native precision is seconds. For 'created in the past 24 hours', use the current instant minus 24 hours through now. Use lastActivityBefore (ISO date or timestamp) for 'no activity since <date>', optionally with sort=lastActivity for stalest first. Different filter types combine with AND; supplied tags match any tag. Reuse the same filters and limit with the opaque nextCursor to continue offset pagination, or omit cursor to refresh. Never page through every contact to export or copy the contact list; point the user to Follow Up Boss's own export. Users without Follow Up Boss export permission get at most 250 contacts per search.
mcp_app_query_people
Inspect Follow Up Boss smart-list names, counts, and only the filter configuration exposed by the readable API. The widget renders the report and supports local filtering plus bounded refresh; do not infer criteria FUB did not return.
mcp_app_query_smart_lists
App-only transport for one bounded Follow Up Boss review action. Requires authentication and the owning workflow's current CSD authorization. Accepts only allowlisted event, workflow, edit-state, count-bucket, and changed-field categories; never prompts, arguments, values, IDs, hashes, credentials, or CRM/customer data. Makes no FUB API calls and does not perform the reviewed product action.
mcp_app_record_review_event
Records feedback the user wrote and submitted in the Share feedback form. Called by that form only. It is not available to the assistant: a record written without the user pressing Submit would attribute words to them that they never wrote.
mcp_app_record_fub_feedback
Purpose: render one final bounded team daily brief after get_team_daily_brief. Source: only the exact team, period, metrics, totals, and coverage returned by that tool plus visibly separate optional annotations. Category: render-only; this makes zero FUB API calls. Selecting one supplied agent persists its exact ID/name and sends one guarded continuation to get_agent_follow_up_leads then render_agent_follow_up_leads. Do not fabricate totals, call create_note, or contact an agent.
mcp_app_render_team_daily_brief
Follow Up Boss ChatGPT Plugin FAQ
How the directory, categories and Discoverability Score work.
Read the methodologyHow do I improve Follow Up Boss's ChatGPT Plugin discoverability?
The levers are the listing surface agents actually read: names, descriptions, keywords, tool metadata, and registry health. Which lever matters depends on where discovery breaks, which is what continuous measurement shows.
What are Follow Up Boss alternatives on ChatGPT?
As of 2026-11-20, Follow Up Boss competes with Agree Zone, ANTOC CRM, BoldTrail, Chat Estelar, Client Keeper Real Estate CRM, Comuro, Connectlinx, Coordnator and 30 more in ChatGPT Real Estate Agent & Broker Software, ranked by public Discoverability Score.
Where is this profile measured?
This profile uses the geography attached to the latest public registry snapshot: US. Locale tags are intentionally omitted.