MCP server

Let Claude Code, Cursor and other AI agents read your reports, fix the bug and close the loop.

FeedbackDots runs a remote Model Context Protocol server. Connect your coding agent to it and you can say "fix the newest high-severity report": the agent reads the report with its console errors, pins and screenshot, changes your code, then marks the report done and comments with what it changed.

URL: https://feedbackdots.com/api/mcp
Transport: Streamable HTTP
Auth: an API key, sent as Authorization: Bearer fd_live_...
Plans: every plan, within the monthly API and MCP request allowance (see plan limits). The feature request tools need Pro.

Create a key under Developers in the dashboard. Name it after the agent, such as "Claude Code", so you can revoke it on its own.

Claude Code

claude mcp add --transport http feedbackdots https://feedbackdots.com/api/mcp \
  --header "Authorization: Bearer fd_live_..."

Cursor

Add this to .cursor/mcp.json in your project, or to ~/.cursor/mcp.json for every project:

{
  "mcpServers": {
    "feedbackdots": {
      "url": "https://feedbackdots.com/api/mcp",
      "headers": { "Authorization": "Bearer fd_live_..." }
    }
  }
}

VS Code

Add this to .vscode/mcp.json. VS Code asks for the key the first time and stores it securely:

{
  "inputs": [
    {
      "type": "promptString",
      "id": "feedbackdots-key",
      "description": "FeedbackDots API key",
      "password": true
    }
  ],
  "servers": {
    "feedbackdots": {
      "type": "http",
      "url": "https://feedbackdots.com/api/mcp",
      "headers": { "Authorization": "Bearer ${input:feedbackdots-key}" }
    }
  }
}

Windsurf and other clients

Any client that supports remote MCP servers with custom headers works with the URL and header above. For clients that only run local servers, bridge with mcp-remote:

{
  "mcpServers": {
    "feedbackdots": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://feedbackdots.com/api/mcp",
        "--header",
        "Authorization: Bearer fd_live_..."
      ]
    }
  }
}

Tools

Tool	What it does
`list_projects`	Lists the workspace's projects and their IDs.
`list_reports`	Lists reports, filtered by project, status, severity or search text.
`get_report`	Returns a report as Markdown with what happened, pins and their CSS selectors, console logs, device, network requests, comments and links, plus the screenshot as an image.
`update_report`	Changes status, severity, title, description, area or archived.
`add_comment`	Comments on a report, for example to say what was fixed.
`create_report`	Files a new report in a project.
`list_feature_requests`	Lists ideas, most votes first.
`create_feature_request`	Adds an idea to a project's board.
`update_feature_request`	Moves an idea to planned, shipped and so on.

Changes made through the MCP server behave like changes in the dashboard: moving a report to Done updates Linear, GitHub or Jira, can email the reporter, and fires webhooks.

Prompts that work well

"List the Triage reports for the storefront project and fix the most severe one. When you're done, move it to In Review and comment with what you changed."
"Get report c3d1… and find the code behind the element in pin 1."
"Which feature requests have the most votes this month? Mark the dark mode one as planned."

Security

The key gives the agent the same access as the API: it can read and change everything in the workspace, but not delete anything. Give each agent its own key and revoke it under Developers when you stop using it.

← Webhooks SDKs →