MCP server

webflutter

Your feedback inbox, readable by your AI tools. Connect Claude Code, Claude Desktop or Cursor to Tracebird and ask things like “what are the critical open bugs?” or “show me the logs of the latest checkout report and fix it”.

Connect

The server speaks MCP over Streamable HTTP at https://api.tracebird.dev/mcp, authenticated with your workspace API token (Settings → API). For Claude Code:

terminal

claude mcp add --transport http tracebird https://api.tracebird.dev/mcp \
  --header "Authorization: Bearer $TRACEBIRD_TOKEN"

Claude Desktop or Cursor (JSON config):

mcp.json

{
  "mcpServers": {
    "tracebird": {
      "type": "http",
      "url": "https://api.tracebird.dev/mcp",
      "headers": { "Authorization": "Bearer <your token>" }
    }
  }
}

⚠ heads upThe token grants read/write access to every report in the workspace — treat it like a password and keep it out of version control.

Tools

option	type	default	description
list_projects	read	—	All projects with untriaged counts.
list_reports	read	—	Recent reports for a project — AI title, severity, status, duplicate grouping. Filter by status.
get_report	read	—	Full detail: tester comment, AI triage, console logs, network, JS errors, navigation, team comments.
update_report	write	—	Change status and/or assignee.
add_comment	write	—	Leave a note for the team.

What it unlocks

Fix from the report — your coding agent reads the suspected cause and the exact failing request, opens the file named in the stack trace, and patches it without you copy-pasting logs.
Triage in chat — “assign every critical checkout bug to Ana and mark the duplicates” is one sentence.
Standup summaries — “what came in this week, by severity?” pulls live numbers, not stale screenshots.

← previousIngest HTTP API next →Security & privacy