Loops MCP server: setup, pricing & agent-readiness

How to connect Loops MCP to Claude Desktop

Add this block to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Cursor and Codex use a similar schema.

{
  "mcpServers": {
    "loops": {
      "command": "npx",
      "args": ["-y", "<server-package-name>"],
      "env": {
        "LOOPS_API_KEY": "your-api-key-here"
      }
    }
  }
}
// Server source: https://loops.so/docs/mcp
// Check the server's README for the exact "command" and required env vars.

Some MCP servers require an API key. Check the source page for setup: loops.so/docs/mcp.

Does Loops MCP cost extra?

Community (free)

The Loops MCP server is community-built — free to run, but you still pay for your Loops subscription.

Community MCPs are normally self-hosted from a GitHub repo. There's no extra license fee.
Your Loops account still needs whatever plan is required to access the data the MCP server queries.
Trade-off: zero MCP cost, but no guarantee it'll keep working when Loops changes their UI or API.

MCP cost

Free (self-hosted)

Subscription required

Yes (your tool plan)

Maintained by

Community

Loops also publishes a REST API

Loops publishes an API, but per-tier gating wasn't extracted automatically.

API docs

—

Auth method

—

SDK languages

—

OpenAPI spec

—

Webhooks

—

Structured outputs

—

See Loops API page

What you can do with Loops from an AI agent

Three shapes of agent workflow. Pick the row that matches how you actually use Claude / Cursor / Codex day to day.

INDIE / SOLO

Skip the dashboard, ask Claude

"What campaigns underperformed this week?" answered in seconds
Quick one-off questions without leaving the editor
Pipe Loops data into your personal notes / docs without writing a script

Effort: 30 seconds — paste the config, restart Claude

AGENCY

Run client analysis at agent speed

Hand the MCP server to your team — every consultant can ask Claude about any client's Loops account
Multi-account workflows without context-switching the Loops UI
Reports you'd normally spend hours on, done before standup

Effort: An afternoon — set up auth per client

IN-HOUSE TEAM

Production agent workflows

Embed in your internal Codex / Cursor agents that triage Loops data daily
Combine with other MCP servers in a multi-step agent chain
Constrain agents to read-only roles to keep production Loops safe

Effort: Days — security review + role config

Limits and gotchas

Fallback: the REST API works for any flow the MCP server doesn't cover.
Rate limits: MCP calls count against your underlying Loops API quota. Burst usage from a curious agent can drain a daily allowance fast.

0/10

Agent-readiness verdict

Not agent-ready: agent workflows need documented APIs and clear integration paths, but Loops has an MCP server available with no public API docs, OpenAPI spec, or SDKs to anchor Claude/Cursor automation. The lack of documented authentication and SDK support makes it difficult to build reliable agent chains without reverse-engineering the integration.

Scored 2026-05-19 by Joonas (TMB).

Loops MCP & API FAQ

Is the Loops MCP server official?+

No — it's a community build, not maintained by Loops. Expect occasional drift and check the README before depending on it for production agents.

Can I use the Loops MCP with Cursor or Codex?+

Yes — MCP is a protocol, not Claude-specific. Any MCP-capable client (Cursor, Codex, Windsurf, Zed) can use the same server config with minor path adjustments.

Sources

Loops official site: https://loops.so
MCP server page: https://loops.so/docs/mcp

Data verified by Joonas on the dates shown. MCP server status auto-rechecked weekly.

Loops MCP Server

What is the Loops MCP server?

What you actually do with the Loops MCP