Today we're launching Quotient's MCP server. Connect Claude — along with Claude Code, Codex, Cursor, and any other MCP-compatible AI app — to your Quotient account, and put it to work creating campaigns, drafting blogs, sending emails, posting to social, reading and writing memory, and pulling analytics. Real marketing work, run from inside the AI tools you already use every day.
The endpoint is https://www.getquotient.ai/mcp. Setup takes a few minutes.
Agent-to-agent, not just tools
Most MCP servers are tool servers. They expose a list of actions — create a record, fetch a row, post a message — and the AI client strings the calls together itself. Quotient's server exposes direct tools too, but it also includes two tools that let the client send a natural-language message to Quotient's marketing agent and get a response back: a2a-send-message and a2a-get-response.
This matters because most marketing work isn't a single API call. Drafting an email broadcast involves brand voice, audience targeting, template structure, and past performance. Setting up a flow involves triggers, conditions, and content for every step. Quotient already knows all of that — it lives in your memory, your segments, your existing templates and flows. When Claude delegates a fuzzy task to Quotient through A2A, it hands the work to an agent that already has the context, instead of trying to learn it on the fly through a chain of tool calls.
Anthropic recommends this kind of design for complex tasks. The docs call out which workflows benefit most from A2A versus the direct tools.
What you can do with it
Together, the direct tools and A2A unlock a wide range of workflows across whichever AI app you happen to be in. A few examples to make this concrete:
Launch a feature from Claude Code. Finish a PR, then ask Claude to summarize what shipped and have Quotient launch it. Quotient produces the campaign, blog draft, email, and social posts, grounded in your brand memory.
Plan a campaign in a Claude conversation. Brainstorm strategy in Claude.ai or Claude Cowork, then hand the brief to Quotient to turn into a real campaign with real tasks.
Pull marketing data into a doc. Writing a board memo or strategy doc in Claude Cowork? Ask Claude to pull last quarter's campaign performance, segment metrics, or list growth directly from Quotient.
Remix existing assets. Ask Claude to pull a welcome email template, rewrite a section for a new product line, and save it back to Quotient. The asset stays in the system of record.
Setup
The endpoint (https://www.getquotient.ai/mcp) is the same for every client. You'll be prompted to sign in to Quotient to authorize access the first time you connect.
Claude
Settings → Connectors → Add custom connector, then enter the endpoint.
Claude Code
claude mcp add --transport http quotient https://www.getquotient.ai/mcp
Then run /mcp in a session to finish authentication.
Codex
codex mcp add quotient --url https://www.getquotient.ai/mcp
Cursor
Settings → MCP Tools → Add custom MCP, then add:
{
"mcpServers": {
"quotient": {
"url": "https://www.getquotient.ai/mcp"
}
}
}Setup for VS Code, Windsurf, Zed, and other MCP clients is in the setup instructions.
Try it
Add the endpoint to your AI client, sign in, and ask Claude to do something useful in your Quotient account. Full setup and tools reference are in the MCP docs.
Don't have a Quotient account yet? Start here.
Blog
Keep up with the latest
from Quotient
Similar Topics