MCP
MCP server
Connect Claude, Cursor, ChatGPT-compatible tooling, or custom clients to OrcaPulse workflows, leads, calls, and integrations through a workspace-scoped MCP server setup.
Overview
The OrcaPulse MCP server is meant for teams that want AI clients to work directly with their OrcaPulse workspace. Instead of building a custom connector for every assistant, you expose approved workflow, lead, call, and integration actions through one MCP-compatible surface.
This page is intentionally practical. It shows what to prepare in the OrcaPulse portal, how to connect an MCP client, what kinds of actions are safe to expose, and which live OrcaPulse links and REST endpoints you can use today while final MCP provisioning is completed for your workspace.
Treat the OrcaPulse MCP server URL like a credential. If your workspace has not yet been provisioned with a dedicated MCP server URL, use the portal and REST API links in this guide first and request MCP enablement before sharing the setup with end users.
Prerequisites
Before planning MCP-style access, make sure the operational pieces inside OrcaPulse are ready.
- Have an OrcaPulse account with access to the live portal at app.orcapulse.ai.
- Confirm the workspace already has the workflows, assistants, lead flows, or CRM integrations you want external systems to work with.
- Have a bearer token or other approved workspace credential for authenticated API access.
- Decide which actions matter most today: read-only analytics, lead creation, workflow lookups, CRM sync, or call retrieval.
- Review the Security documentation before using MCP in production environments.
Current state
OrcaPulse does not currently expose a public self-serve MCP server endpoint that users can provision directly from this portal. So this page should be read as a practical guide to the real OrcaPulse surfaces that an MCP server would eventually sit on top of, not as a live endpoint setup screen.
The product surfaces that exist today are the OrcaPulse portal, integrations, workflows, assistants, lead handling, and the public REST API. Those are the real entrypoints users can work with right now.
What users can do today
- Log in to the OrcaPulse portal and verify the correct workspace.
- Create or review workflows, assistants, leads, and integrations in the current product.
- Use the REST API for automation around workflows, calls, leads, and integrations.
- Treat any MCP server discussion as support-assisted or future-facing unless OrcaPulse provides a real endpoint to your workspace.
Do not hardcode a guessed MCP endpoint. If you need real automation today, the live REST base is https://api.orcapulse.ai.
Current surface
The most useful OrcaPulse MCP-style surface would be built on top of actions the product already supports today. That means the table below reflects the real underlying product and API surface, not a live self-serve MCP endpoint.
| MCP surface | What users do | Related OrcaPulse route |
|---|---|---|
| Workflow lookup | List workflows, inspect status, and understand routing or automation paths. | GET /workflows, GET /workflows/:id, PATCH /workflows/:id/status |
| Lead intake | Create new leads from chat, forms, or operator requests without leaving the AI client. | GET /leads/all, POST /leads/manual |
| Call retrieval | Find calls, open transcripts, and summarize outcomes. | GET /calls, GET /calls/:id, GET /calls/:id/transcript |
| CRM sync | Push qualified lead data into connected external systems. | POST /integrations/:id/sync-lead, POST /integrations/:id/test |
| Messaging readiness | Inspect WhatsApp, SMS, and email assets before downstream automations run. | GET /whatsapp, GET /sms-templates, GET /email-templates |
Recommended real-world starting point: use workflow lookup, lead creation, call retrieval, and integration sync through the current OrcaPulse API before treating them as MCP-style tools.
Example usage
Even when the MCP layer is your main interface, these live OrcaPulse REST examples are helpful for testing the exact business actions the MCP server should wrap.
List workflows
curl -G https://api.orcapulse.ai/workflows \ -H "Authorization: Bearer YOUR_ORCAPULSE_BEARER_TOKEN"Create a manual lead
curl https://api.orcapulse.ai/leads/manual \ -X POST \ -H "Authorization: Bearer YOUR_ORCAPULSE_BEARER_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "name": "Ava Thompson", "phone": "+14155550123", "email": "ava@example.com", "source": "mcp-test" }'List calls
curl -G https://api.orcapulse.ai/calls \ -H "Authorization: Bearer YOUR_ORCAPULSE_BEARER_TOKEN"Sync a lead to an integration
curl https://api.orcapulse.ai/integrations/integration_001/sync-lead \ -X POST \ -H "Authorization: Bearer YOUR_ORCAPULSE_BEARER_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "leadId": "lead_001" }'Good user prompts
- Operations: "Show me the active OrcaPulse workflows and highlight any that are paused."
- Sales: "Create a lead for Sarah from Acme and tell me which workflow should process it."
- Support: "Fetch the last call transcript for this phone number and summarize the unresolved issue."
- RevOps: "Push this qualified lead into the connected CRM integration and confirm the sync result."
Best practices
- Document only real product paths: if a public MCP endpoint is not present, guide users through portal and API flows instead.
- Start with supported actions: workflows, leads, calls, and integrations are the strongest current surface.
- Keep access narrow: any future MCP-style layer should stay scoped and approval-aware.
- Plan a fallback path: users should always be able to continue through the OrcaPulse portal or direct API.