MCP Tools Reference

CRM Factory's MCP server exposes 14 tools across three categories: CRM operations, ERP queries, and knowledge search.

---

CRM Tools

search_crm

Search CRM records by keyword across multiple entity types.

Returns: { query, results: [{ id, type, title, subtitle }] }

Performs case-insensitive substring matching across name, email, subject, and other text fields. Results are merged and capped at limit × 3 total.

---

get_record

Retrieve a single CRM record with all related entities.

Returns vary by type:

• account: { record, contacts, opportunities, cases, recentActivities }
• contact: { record, account, opportunities, cases, recentActivities }
• opportunity: { record, account, contact, products, recentActivities }
• case: { record, account, contact, comments, recentActivities }
• lead: { record }

---

list_records

List CRM records with optional filtering.

Returns: { type, records: [...] }

---

create_record

Create a new CRM record. Supported types: opportunity, case, activity.

Opportunity fields: name (required), stage (default "Discovery"), account_id, contact_id, amount, currency, probability, expected_close_date, pipeline_id

Case fields: subject (required), description, account_id, contact_id, priority (default "medium"), status (default "new"), category

Activity fields: activity_type (required: email/call/meeting/task/note), subject (required), body, due_at, related_to (array of { type, id })

The authenticated user is automatically set as the owner.

---

update_record

Update fields on an existing CRM record.

Editable fields vary by type. For opportunities: name, stage, amount, probability, expected_close_date, loss_reason, won. For cases: subject, description, status, priority, category.

---

get_pipeline_summary

Get aggregated pipeline metrics with opportunity count and total value per stage.

Returns: { pipeline_id, stages: [{ stage, opportunityCount, totalAmount }] }

---

draft_email

Draft a customer-facing email using CRM context. Uses AI to generate a polished email based on the contact's account, open opportunities, and recent activity.

Returns: { to, subjectSuggestion, draft, context: { contactId, accountId, opportunityIds } }

---

get_call_history

Get recent call activity for a contact or account.

*At least one of contact_id or account_id is required.

Returns: { calls: [{ id, subject, created_at, completed_at, duration, direction, call_sid, ai_summary }] }

---

search_call_transcripts

Search call transcripts by keyword.

Returns: { query, results: [{ id, subject, created_at, transcript }] }

---

get_call_analysis

Get full AI post-call analysis for a single call.

Returns: { id, subject, created_at, completed_at, call_sid, transcript, analysis }

The analysis object contains AI-generated insights including summary, action items, and sentiment.

---

ERP Tools

These tools require an active ERP integration connection (Infor SX, SAP, NetSuite, or Dynamics).

erp_lookup

Lookup ERP data by type.

• account: { account_id?, email?, name?, erp_customer_id? } — Looks up from local synced data
• product: { sku?, query?, limit? } — Looks up from local synced data
• order: { order_number } — Live ERP query
• order_history: { account_id?, customer_id?, limit? } — Live ERP query
• pricing: { customer_id, sku?, items?: [{ sku, qty? }] } — Live ERP query
• customer_kpis: { customer_id } — Live ERP query

---

erp_check_inventory

Check live ERP inventory levels.

Returns: { skus, inventory: [{ sku, warehouseId, quantityOnHand, quantityAvailable, quantityOnOrder }] }

---

erp_live_query

Generic live ERP query dispatcher for advanced use cases.

---

Knowledge Tools

search_knowledge

Semantic vector search across the embedded knowledge base.

Returns: { query, results: [{ id, sourceType, sourceId, parentType, parentId, content, metadata, similarity }] }

Uses pgvector cosine similarity to find the most semantically relevant knowledge chunks. Higher min_similarity values return fewer but more relevant results.