Quickstart

Two minutes from "I have a paid account" to "Claude is querying my research."

1. Generate an API key

In the app: Settings → API Keys → Generate API Key.

You'll see the full ue_live_… key once at creation — copy it. After that, only the last four characters are recoverable.

If you lose a key, generate a new one and revoke the old.

2. Connect a client

Pick the one you use. All four expose the same 13 tools.

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows). If the file doesn't exist, create it.

{
  "mcpServers": {
    "userevaluation": {
      "type": "http",
      "url": "https://api.userevaluation.com/mcp",
      "headers": {
        "Authorization": "Bearer ue_live_..."
      }
    }
  }
}

Restart Claude Desktop. The 🔨 icon at the bottom of the conversation should now list the UE tools.

If you'd rather run it locally (no outbound HTTPS, useful for IT-restricted environments):

{
  "mcpServers": {
    "userevaluation": {
      "command": "npx",
      "args": ["-y", "@userevaluation/mcp"],
      "env": { "UE_API_KEY": "ue_live_..." }
    }
  }
}

Claude.ai (web)

Settings → Connectors → Add custom connector.

Field	Value
Name	User Evaluation
URL	https://api.userevaluation.com/mcp
Auth	Bearer token: ue_live_...

Available on Pro / Max / Team / Enterprise plans.

Cursor

Settings → MCP → Add Server.

{
  "mcpServers": {
    "userevaluation": {
      "type": "http",
      "url": "https://api.userevaluation.com/mcp",
      "headers": { "Authorization": "Bearer ue_live_..." }
    }
  }
}

Claude Code (CLI)

claude mcp add-json userevaluation '{
  "type": "http",
  "url": "https://api.userevaluation.com/mcp",
  "headers": { "Authorization": "Bearer ue_live_..." }
}'

Or for the local stdio variant:

claude mcp add-json userevaluation '{
  "command": "npx",
  "args": ["-y", "@userevaluation/mcp"],
  "env": { "UE_API_KEY": "ue_live_..." }
}'

ChatGPT

Settings → Apps & Connectors → Add custom connector. Paste https://api.userevaluation.com/mcp with bearer auth.

3. Verify

In the conversation, ask:

List my User Evaluation projects.

The model should call the list_projects tool and return a table. If you see a plan_required error, your account is on Free — upgrade to use the API and MCP.

4. Try a real workflow

Find the file in my "Onboarding study" project that mentions step 2 the most, and create a pain-point card with the participant's exact quote.

Behind the scenes the model will:

1. list_projects to find the study
2. list_project_files to see what's available
3. get_transcript on each file to scan
4. create_card to write the insight back into UE

Troubleshooting

Symptom	Likely cause	Fix
unauthorized on every call	Bad / expired bearer	Generate a fresh key in Settings
plan_required	Free account	Upgrade to Basic / Standard / Plus
forbidden	Signed in as a participant	Use a researcher account
rate_limited	Burst too fast	Backoff per Retry-After; or upgrade plan
MCP tools don't show up in the client	Config path / format error	Check the client's MCP-server log; the file must be valid JSON

→ Tools reference — full list of what's exposed and what each does.