MCP

The Whippy MCP server lets an AI assistant work in your Whippy workspace for you. Once connected, you can ask it to find contacts, read conversations, build segments, review campaigns, or send a message, and it calls Whippy on your behalf.

MCP (Model Context Protocol) is the open standard that AI assistants such as Claude, ChatGPT, and Cursor use to connect to outside tools. Whippy hosts the server, so there is nothing to install or run.

Server URLhttps://api.whippy.co/mcp
TransportStreamable HTTP (remote)
AuthenticationOAuth 2.1 with PKCE. You sign in with your Whippy account; no API key is needed
Tools256, one per Whippy public API endpoint. See the tool reference
Rate limit10 requests per second for each user in each workspace. See Rate limits
📘

Two different MCP servers

https://api.whippy.co/mcp works with your Whippy data (contacts, messages, campaigns, and so on). https://docs.whippy.ai/mcp is a separate server that only searches this documentation site. Use the first to get work done in Whippy and the second to help an AI write code against the Whippy API.

Connect a client

Pick your assistant. Each guide takes a few minutes.

ClientGuide
Claude (claude.ai, Claude Desktop, Claude mobile)Connect Claude
Claude CodeConnect Claude Code
ChatGPTConnect ChatGPT
CursorConnect Cursor
DevinConnect Devin
Grok (grok.com, Grok Build)Connect Grok
VS Code (GitHub Copilot)Connect VS Code
WindsurfConnect Windsurf
Another MCP clientConnect any MCP client

What the assistant can do

The assistant gets the same access you have. It acts as you, in the workspace you choose when you sign in, and it can only do what your Whippy role allows. Every tool call goes through the same checks as a call to the Whippy public API.

Tools cover the whole public API:

  • Contacts and data: search and filter contacts (including with WQL), manage tags, lists, segments, custom objects, and communication preferences.
  • Messaging: search conversations and messages, send SMS and email, add notes, and manage templates.
  • Campaigns and sequences: list and inspect campaigns, enroll contacts in sequences, and review sequence runs.
  • AI Agents: list agents, review calls and runs, manage knowledge bases and tests.
  • Workflows: list workflows, apps, and connections, and review runs.
  • Settings: teams, users, channels, schedules, invites, and audit logs.

See the tool reference for every tool.

Example prompts

These prompts name the tools the assistant usually picks, so you can see what happens behind the scenes.

Find a contact and their history

Find the contact with phone number +1 555 010 2233 and summarize their last five conversations.

The assistant calls search_contacts to find the contact, then search_conversations and show_conversation to read the threads.

Build an audience

How many contacts are tagged "Spring Hiring" and have not replied in the last 30 days? Save them as a segment called "Spring Hiring – no reply".

The assistant calls list_tags to find the tag, filter_contacts with a WQL filter to count matching contacts, and create_segment to save the audience.

Review a campaign

List the campaigns we sent last week and who each one went to.

The assistant calls list_campaigns, then show_campaign and list_campaign_contacts for each campaign.

Send a message, with confirmation

Text Jordan Lee from our main number: "Your interview is confirmed for 2pm tomorrow."

The assistant calls search_contacts and list_channels to find the contact and the channel, then send_sms. Most clients ask you to approve tools that change data before they run. Keep that setting on for tools that send messages or delete data.

Next steps