Model Context Protocol (MCP) is an open protocol that lets an external client call typed tools on a server over a standard transport. Dureach implements an inbound MCP server: your client connects to the workspace, lists granted dureach.* tools, and calls them with checked arguments and workspace quotas.
When does an MCP connection fit a GTM workflow?
Use MCP when a team wants research, list review, or draft preparation inside an existing MCP-capable client instead of copying data between tools. Keep the browser app and REST API for bulk administration, webhook automation, and workspace settings. MCP fits interactive research and review; it does not replace ownership of eligibility, consent, and send approval.
Signals -> Leads -> Enrichment -> Decision -> Outreach -> PipelineTreat every tool result as candidate data. Confirm identity, check consent and suppression, and approve customer-facing work before it leaves the workspace.
How is inbound MCP different from other AI connections?
Tool results are transferred to the external client and may be processed or retained by its provider. Review the provider’s data handling, retention, and access terms before connecting. Request only the fields needed for the job; keep contact details and customer records out of shared conversations, screenshots, and logs.
What should owners set up first?
Define the workspace, the permitted data sources, the roles that may approve R2/R3 work, and the daily search budget. Grant the smallest scopes that answer the job: companies:read and trade:read for research, contacts:read for selective lookup, leads:create and campaigns:create for drafting, emails:send or sms:send only where a standing send policy exists. Connection steps are in Connect an AI client via MCP.
What are common questions about MCP for GTM?
What is Model Context Protocol?
It is a standard way for a client to discover tools (tools/list) and call them (tools/call) with typed arguments and structured results.
What does Dureach expose over MCP?
Fourteen dureach.* tools for workspace context, company/trade/contact reads, read-only SQL on two views, lead and campaign reads and drafts, and policy-gated email/SMS sends. See the MCP tool reference.
Does MCP move workspace data to the client vendor?
Tool results are transferred to the external client and may be processed or retained by its provider. Review the provider’s data handling, retention, and access terms before connecting. Request only the fields needed for the job; keep contact details and customer records out of shared conversations, screenshots, and logs.
Who approves outreach started over MCP?
A named workspace owner. Drafts and sends need the same eligibility, claim, consent, suppression, jurisdiction, and licensing checks as app-created outreach.
Where should teams start?
Start with read-only research, review govern AI outreach with MCP approvals, then add writes behind dry runs and approvals.
How does MCP relate to the REST API and webhooks?
Use MCP for interactive tool calls and the documented REST API for supported server integrations. Webhooks notify your systems about supported events. Confirm the supported API contract before building an integration; do not infer public endpoints from browser or internal routes.