brosh_create
ChatGPTCreate one or more CRM records. Pass an array of field objects. Schema is table-aware and includes common fields per table to improve first-call accuracy. Example: create a new lead/contact/opportunity from prospecting output.
Business
Overview
BROSH MCP brings your CRM, sales, marketing, projects, documents, and business operations directly into AI assistants through the Model Context Protocol (MCP). Instead of switching between applications, AI agents can securely access, analyze, update, manage settings, and automate workflows using live data from your BROSH workspace. Designed for sales teams, consultants, service providers, agencies, and growing businesses, BROSH MCP enables AI to work with real business context—including leads, contacts, opportunities, projects, tasks, invoices, contracts, support tickets, and marketing activities. Users can ask questions in natural language, generate insights, automate repetitive processes, and execute CRM actions directly from their AI environment. Key capabilities include: Secure MCP-based access to CRM and business data AI-powered querying, reporting, and data analysis Lead, contact, account, and opportunity management Project, task, and workflow automation Marketing automation and customer engagement workflows Document generation, e-signatures, quotes, and invoicing Custom business objects, fields, dashboards, and reporting Real-time interaction between AI agents and operational systems BROSH MCP transforms AI assistants from simple chat interfaces into business operators capable of understanding your organization, retrieving information, updating records, and helping teams move faster with less manual work. Built on the open MCP standard, it provides a secure bridge between AI and the systems your business relies on every day.
Tools
brosh_create
ChatGPTCreate one or more CRM records. Pass an array of field objects. Schema is table-aware and includes common fields per table to improve first-call accuracy. Example: create a new lead/contact/opportunity from prospecting output.
brosh_create
ChatGPTCreate one or more CRM records. Pass an array of field objects. Schema is table-aware and includes common fields per table to improve first-call accuracy. Example: create a new lead/contact/opportunity from prospecting output.
brosh_delete
ChatGPTDelete CRM records by ID. Deleted records move to recycle bin for 30 days. This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Example: remove duplicate test records after cleanup.
brosh_delete
ChatGPTDelete CRM records by ID. Deleted records move to recycle bin for 30 days. This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Example: remove duplicate test records after cleanup.
brosh_delete
ChatGPTDelete CRM records by ID. Deleted records move to recycle bin for 30 days. This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Example: remove duplicate test records after cleanup.
brosh_find
ChatGPTSearch CRM records with optional filtering, sorting, and pagination. Standard sales opportunities are stored in opportunities with type="Opportunity". Quotes and invoices are also stored in the opportunities table as types/subtypes, and quote/invoice lines are stored in opportunity_products. In opportunities, common fields include name, stage, amount, type, account, contact, owner, close_date, created_date, last_modified_date, currency, tax, discount, and sub_total. Type values are Opportunity, Quote, Invoice. Stage values include Prospecting, Qualification, Needs Analysis, Value Proposition, Id. Decision Makers, Perception Analysis, Proposal/Price Quote, Negotiation/Review, Closed Won, Closed Lost. In contacts, common fields include name, first_name, last_name, email, phone, mobile_phone, account, owner, type, country, town, title, department, industry, linkedin, email_opt_out, address, created_date, and last_modified_date. Contact type values are typically lead/customer (and some views may include contact). In accounts, common fields include id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Account type values include Lead, Account, and Competitor. In products, common fields include name, productcode, description, price, list_price, currency, link, created_date, and last_modified_date. Product currency values include USD, EUR, GBP, AED, AUD, BRL, ILS, HKD, JPY, NZD, PLN, QAR, and RUB. In templates, common fields include id, name, subject, description, type, sub_type, sub_title, body, created_date, last_modified_date, and last_modified_by. Template type values include email, web + pdf, web + pdf + sign, and web form. Template sub_type values may include Marketing, Finance, Sales, Sales & Finance, and All. Activity stores emails, e-signatures on documents, tasks, and events linked to multiple related objects, and is also used for project tasks. In activity, core fields include name (subject), type, status, owner, person, account, opportunities, project, start_date, end_date, progress, last_modified_date, last_modified_by, created_date, ticket, campaign, template, from, to, cc, message_status, message_additional_info, file, and created_by. In tickets, common fields include name (subject), description, contact, email, status, priority, origin, owner, account, created_date, last_modified_date, and last_modified_by. Ticket status values include New, In Work, Closed. Ticket priority values include High, Medium, Low. Ticket origin values include Phone, Web, Call. Use filter.search for free-text on the name field, filter.where for field-level conditions (plain value or {operator, value}), and sort for ordering results. Example: find opportunities where type="Opportunity" sorted by last_modified_date DESC, or filter opportunities by type="Quote" / type="Invoice", or query activity by related object/project fields, or query tickets by status/priority/origin/contact/account.
brosh_find
ChatGPTSearch CRM records with optional filtering, sorting, and pagination. Standard sales opportunities are stored in opportunities with type="Opportunity". Quotes and invoices are also stored in the opportunities table as types/subtypes, and quote/invoice lines are stored in opportunity_products. In opportunities, common fields include name, stage, amount, type, account, contact, owner, close_date, created_date, last_modified_date, currency, tax, discount, and sub_total. Type values are Opportunity, Quote, Invoice. Stage values include Prospecting, Qualification, Needs Analysis, Value Proposition, Id. Decision Makers, Perception Analysis, Proposal/Price Quote, Negotiation/Review, Closed Won, Closed Lost. In contacts, common fields include name, first_name, last_name, email, phone, mobile_phone, account, owner, type, country, town, title, department, industry, linkedin, email_opt_out, address, created_date, and last_modified_date. Contact type values are typically lead/customer (and some views may include contact). In accounts, common fields include id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Account type values include Lead, Account, and Competitor. In products, common fields include name, productcode, description, price, list_price, currency, link, created_date, and last_modified_date. Product currency values include USD, EUR, GBP, AED, AUD, BRL, ILS, HKD, JPY, NZD, PLN, QAR, and RUB. In templates, common fields include id, name, subject, description, type, sub_type, sub_title, body, created_date, last_modified_date, and last_modified_by. Template type values include email, web + pdf, web + pdf + sign, and web form. Template sub_type values may include Marketing, Finance, Sales, Sales & Finance, and All. Activity stores emails, e-signatures on documents, tasks, and events linked to multiple related objects, and is also used for project tasks. In activity, core fields include name (subject), type, status, owner, person, account, opportunities, project, start_date, end_date, progress, last_modified_date, last_modified_by, created_date, ticket, campaign, template, from, to, cc, message_status, message_additional_info, file, and created_by. In tickets, common fields include name (subject), description, contact, email, status, priority, origin, owner, account, created_date, last_modified_date, and last_modified_by. Ticket status values include New, In Work, Closed. Ticket priority values include High, Medium, Low. Ticket origin values include Phone, Web, Call. Use filter.search for free-text on the name field, filter.where for field-level conditions (plain value or {operator, value}), and sort for ordering results. Example: find opportunities where type="Opportunity" sorted by last_modified_date DESC, or filter opportunities by type="Quote" / type="Invoice", or query activity by related object/project fields, or query tickets by status/priority/origin/contact/account.
brosh_find
ChatGPTSearch CRM records with optional filtering, sorting, and pagination. Standard sales opportunities are stored in opportunities with type="Opportunity". Quotes and invoices are also stored in the opportunities table as types/subtypes, and quote/invoice lines are stored in opportunity_products. In opportunities, common fields include name, stage, amount, type, account, contact, owner, close_date, created_date, last_modified_date, currency, tax, discount, and sub_total. Type values are Opportunity, Quote, Invoice. Stage values include Prospecting, Qualification, Needs Analysis, Value Proposition, Id. Decision Makers, Perception Analysis, Proposal/Price Quote, Negotiation/Review, Closed Won, Closed Lost. In contacts, common fields include name, first_name, last_name, email, phone, mobile_phone, account, owner, type, country, town, title, department, industry, linkedin, email_opt_out, address, created_date, and last_modified_date. Contact type values are typically lead/customer (and some views may include contact). In accounts, common fields include id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Account type values include Lead, Account, and Competitor. In products, common fields include name, productcode, description, price, list_price, currency, link, created_date, and last_modified_date. Product currency values include USD, EUR, GBP, AED, AUD, BRL, ILS, HKD, JPY, NZD, PLN, QAR, and RUB. In templates, common fields include id, name, subject, description, type, sub_type, sub_title, body, created_date, last_modified_date, and last_modified_by. Template type values include email, web + pdf, web + pdf + sign, and web form. Template sub_type values may include Marketing, Finance, Sales, Sales & Finance, and All. Activity stores emails, e-signatures on documents, tasks, and events linked to multiple related objects, and is also used for project tasks. In activity, core fields include name (subject), type, status, owner, person, account, opportunities, project, start_date, end_date, progress, last_modified_date, last_modified_by, created_date, ticket, campaign, template, from, to, cc, message_status, message_additional_info, file, and created_by. In tickets, common fields include name (subject), description, contact, email, status, priority, origin, owner, account, created_date, last_modified_date, and last_modified_by. Ticket status values include New, In Work, Closed. Ticket priority values include High, Medium, Low. Ticket origin values include Phone, Web, Call. Use filter.search for free-text on the name field, filter.where for field-level conditions (plain value or {operator, value}), and sort for ordering results. Example: find opportunities where type="Opportunity" sorted by last_modified_date DESC, or filter opportunities by type="Quote" / type="Invoice", or query activity by related object/project fields, or query tickets by status/priority/origin/contact/account.
brosh_get
ChatGPTRetrieve specific CRM records by their numeric IDs. Use table="accounts" for company records and account-level ownership/profile fields such as id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Use table="products" for product catalog records (pricing, SKU/code, currency, links). Use table="templates" for email/document/web templates and their body/subject/type metadata. Use table="opportunities" for Opportunity/quote/invoice headers and table="opportunity_products" for quote/invoice lines. Use table="tickets" for support/helpdesk records (subject/name, description, contact, email, status, priority, origin, owner, account, created/modified timestamps). Use table="activity" for email logs, e-signature activities, tasks/events, and project tasks, linked by fields like person/account/opportunities/project/ticket/campaign. Example: fetch contacts by IDs to enrich or validate before update.
brosh_get
ChatGPTRetrieve specific CRM records by their numeric IDs. Use table="accounts" for company records and account-level ownership/profile fields such as id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Use table="products" for product catalog records (pricing, SKU/code, currency, links). Use table="templates" for email/document/web templates and their body/subject/type metadata. Use table="opportunities" for Opportunity/quote/invoice headers and table="opportunity_products" for quote/invoice lines. Use table="tickets" for support/helpdesk records (subject/name, description, contact, email, status, priority, origin, owner, account, created/modified timestamps). Use table="activity" for email logs, e-signature activities, tasks/events, and project tasks, linked by fields like person/account/opportunities/project/ticket/campaign. Example: fetch contacts by IDs to enrich or validate before update.
brosh_get
ChatGPTRetrieve specific CRM records by their numeric IDs. Use table="accounts" for company records and account-level ownership/profile fields such as id, name, phone, website, address, country, owner, type, industry, company_size, headquarters, founded, linkedin, created_date, last_modified_date, last_modified_by, created_by, number_of_contacts-exp, and number_of_users-exp. Use table="products" for product catalog records (pricing, SKU/code, currency, links). Use table="templates" for email/document/web templates and their body/subject/type metadata. Use table="opportunities" for Opportunity/quote/invoice headers and table="opportunity_products" for quote/invoice lines. Use table="tickets" for support/helpdesk records (subject/name, description, contact, email, status, priority, origin, owner, account, created/modified timestamps). Use table="activity" for email logs, e-signature activities, tasks/events, and project tasks, linked by fields like person/account/opportunities/project/ticket/campaign. Example: fetch contacts by IDs to enrich or validate before update.
brosh_me
ChatGPTGet the current authenticated BROSH CRM user profile. Example: call this first to confirm auth and inspect user role/account context.
brosh_me
ChatGPTGet the current authenticated BROSH CRM user profile. Example: call this first to confirm auth and inspect user role/account context.
brosh_me
ChatGPTGet the current authenticated BROSH CRM user profile. Example: call this first to confirm auth and inspect user role/account context.
brosh_modify_field
ChatGPTModify or create a CRM table field using BROSH modifyField API. Calls /api/oauth2/modifyfield/:source/:tableName and expects oldField + newField objects (or legacy xFields as [oldField, newField]). Supported format3 values: VARCHAR, DATE, DATETIME, DECIMAL, INT, BLOB, TEXT, MEDIUMTEXT. Validation mirrors BROSH constraints for type/length/decimals before sending.
brosh_modify_field
ChatGPTModify or create a CRM table field using BROSH modifyField API. Calls /api/oauth2/modifyfield/:source/:tableName and expects oldField + newField objects (or legacy xFields as [oldField, newField]). Supported format3 values: VARCHAR, DATE, DATETIME, DECIMAL, INT, BLOB, TEXT, MEDIUMTEXT. Validation mirrors BROSH constraints for type/length/decimals before sending.
brosh_modify_field
ChatGPTModify or create a CRM table field using BROSH modifyField API. Calls /api/oauth2/modifyfield/:source/:tableName and expects oldField + newField objects (or legacy xFields as [oldField, newField]). Supported format3 values: VARCHAR, DATE, DATETIME, DECIMAL, INT, BLOB, TEXT, MEDIUMTEXT. Validation mirrors BROSH constraints for type/length/decimals before sending.
brosh_send_email
ChatGPTSend emails using BROSH CRM email engine. Supports template lookup by templateid, sender lookup by fromEmailSet, and direct template payloads. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. The upstream API validates sender permissions and required email content. Example: send a campaign email to filtered contact rows using a template and approved sender. Tip: if you process one row at a time, template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_send_email
ChatGPTSend emails using BROSH CRM email engine. Supports template lookup by templateid, sender lookup by fromEmailSet, and direct template payloads. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. The upstream API validates sender permissions and required email content. Example: send a campaign email to filtered contact rows using a template and approved sender. Tip: if you process one row at a time, template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_send_email
ChatGPTSend emails using BROSH CRM email engine. Supports template lookup by templateid, sender lookup by fromEmailSet, and direct template payloads. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. The upstream API validates sender permissions and required email content. Example: send a campaign email to filtered contact rows using a template and approved sender. Tip: if you process one row at a time, template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_send_message_single
ChatGPTSend emails one-by-one with a custom message per row. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Each row can include message/body and subject, and MCP will call BROSH sendEmail per recipient. Use this when AI generates a unique message for each contact/account. Tip: this method processes one row at a time, and template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_send_message_single
ChatGPTSend emails one-by-one with a custom message per row. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Each row can include message/body and subject, and MCP will call BROSH sendEmail per recipient. Use this when AI generates a unique message for each contact/account. Tip: this method processes one row at a time, and template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_send_message_single
ChatGPTSend emails one-by-one with a custom message per row. Before sending, MCP clients must query the email_settings table first to resolve the sender ID (fromEmailSet) and sender record (fromEmailSets). This tool requires explicit two-step confirmation: first call returns a preview + confirmationToken; second call must include confirm=true and the token. Each row can include message/body and subject, and MCP will call BROSH sendEmail per recipient. Use this when AI generates a unique message for each contact/account. Tip: this method processes one row at a time, and template.body can contain the exact AI-generated personalized message for that recipient, without parameters to replace.
brosh_update
ChatGPTUpdate one or more CRM records. Each object must include an "id" field. Schema is table-aware and includes common updatable fields per table to improve first-call accuracy. Example: update missing fields after AI enrichment.
brosh_update
ChatGPTUpdate one or more CRM records. Each object must include an "id" field. Schema is table-aware and includes common updatable fields per table to improve first-call accuracy. Example: update missing fields after AI enrichment.
brosh_update
ChatGPTUpdate one or more CRM records. Each object must include an "id" field. Schema is table-aware and includes common updatable fields per table to improve first-call accuracy. Example: update missing fields after AI enrichment.
Capabilities
Writes
