MCP — Moevox API

Connect Moevox to your AI workflow (Claude Desktop, Cursor, ChatGPT, and other MCP clients) with a one-line config — no code, no polling logic to write.

What you get

The MCP endpoint exposes the same research pipeline as three tools. Your AI assistant decides when to call them — you just talk to it.

Tool	What it does
`create_research_report`	Submit a report request asynchronously and get a `request_id` back immediately.
`get_report_status`	Poll the progress of a submitted request until it completes.
`check_credits`	Check your remaining credit balance before starting.

The AI handles the whole async flow for you: it checks credits, submits the report, polls until done, and summarizes the findings. You never see a request_id unless something goes wrong.

Setup — 3 steps

Create an API key — open your API keys page in the research workspace and create one. You'll get a key that looks like moe_sk_live_....
Copy this config and replace moe_sk_live_... with your own key from step 1 (everything else stays the same):
Paste it into your MCP client (see where to paste below).

mcp config

{
  "mcpServers": {
    "moevox": {
      "url": "https://api.moevox.com/api/v1/mcp",
      "headers": { "Authorization": "Bearer moe_sk_live_..." }
    }
  }
}

That is the whole setup. The API key authenticates every call, and credits are consumed only when a report actually samples respondents.

Try it

Once connected, you can simply ask in plain language — in any language:

“Which of these pricing plans should a small SaaS launch first: Monthly $29, Monthly $49, or Annual $299?”

The assistant checks your credits, runs the report, and comes back with the winner, confidence level, key risks, and a link to the public report.

Supported clients

The same config works everywhere. Pick your client:

Claude Code (CLI) — one line, then replace the key:

claude mcp add moevox --transport http https://api.moevox.com/api/v1/mcp --header "Authorization: Bearer moe_sk_live_..."

Claude Desktop — add the config to claude_desktop_config.json
Cursor — add it in Settings → MCP, or to .cursor/mcp.json
ChatGPT — connect a remote MCP server with the URL above
Any MCP client that supports remote (Streamable HTTP) servers

The config block is identical everywhere — you only ever change the moe_sk_live_... to your own key.

Where to paste the config

Every client stores MCP servers differently, but the config block is the same everywhere. Look for the client's MCP settings:

Claude Desktop: Settings → Developer → Edit Config
Cursor: Settings → MCP → Add new MCP server
Windsurf: Settings → MCP Servers → Add

Prefer not to configure a client? Copy the AI Prompt into any LLM chat instead — it works with zero setup.