ReportPortal MCP Server – Test Analytics for AI
ReportPortal MCP Server lets MCP-compatible AI clients query ReportPortal launches, failures, analysis data, and Test Case Management resources. It is useful for QA, release, and engineering workflows that need conversational access to recent test runs and test design assets.
Overview
ReportPortal MCP Server is the official Model Context Protocol server from the ReportPortal project. It connects AI assistants such as Claude Desktop, Cursor, VS Code Copilot, JetBrains Copilot, and Claude Code to a ReportPortal instance so they can retrieve testing and quality data through structured MCP tools instead of relying on screenshots, manual navigation, or ad hoc scripts.
What the MCP server enables
The server exposes ReportPortal data to AI agents for common test-reporting questions and QA workflows. Agents can ask for recent launches, failed tests, failure summaries, error details, and other launch analysis context. Starting with the 2.x line, the server also integrates with ReportPortal Test Case Management so agents can browse and manage test cases, folders, milestones, and test plans from the same conversational interface.
Practical uses include asking an assistant what failed in the latest run, listing the most recent launches, summarizing recurring failures, browsing test cases in a specific folder, or preparing a test plan for a sprint milestone. Access is still governed by the ReportPortal API token and the project key supplied to the server or request.
When to use it
Use this MCP server when teams already store automated test execution, quality reporting, or TMS data in ReportPortal and want AI-assisted investigation or planning. It is well suited for QA engineers, release managers, and developers who need quick test insights while working in an MCP-capable coding or chat tool.
Connection and authentication
ReportPortal supports local stdio operation and remotely deployed HTTP/SSE operation. For local use, the documented Docker configuration runs reportportal/mcp-server with RP_HOST, RP_API_TOKEN, and optionally RP_PROJECT. Pre-built binaries can also run in stdio mode with the stdio argument. For remote use, deploy the server with MCP_MODE=http; clients connect to /mcp or /api/mcp and pass Authorization: Bearer <token> plus optional X-Project per request. The API token must come from the user's ReportPortal profile and must never be committed to source control.
Key considerations
The server does not provide a shared public hosted endpoint; HTTP mode requires a deployed server. RP_HOST must include the protocol and point to an accessible ReportPortal instance. In stdio mode, RP_PROJECT can set a default project, while HTTP mode ignores RP_PROJECT and expects project context through the X-Project header. Invalid tokens produce authentication errors, and valid tokens without project permissions can still fail with authorization errors.
Supported Transports
stdio
Command: docker
Args:
run-i--rm-eRP_API_TOKEN-eRP_HOST-eRP_PROJECTreportportal/mcp-server
streamable_http
URL: http://localhost:8080/mcp
sse
URL: http://localhost:8080/mcp
Frequently Asked Questions
- When should an AI agent use the ReportPortal MCP Server?
- Use it when the agent needs live ReportPortal context, such as recent launches, failed tests, failure counts, error details, launch analysis, or Test Case Management data for QA planning and release investigation.
- What does this MCP server add to an AI agent's capabilities?
- It gives the agent structured access to ReportPortal testing and TMS data through MCP tools, so the agent can answer test-reporting questions without manually navigating the ReportPortal UI or relying on static model knowledge.
- What can the AI agent access or manage through this server?
- The documented server supports launch and failure-oriented queries and, in the 2.x line, ReportPortal Test Case Management resources such as test cases, folders, milestones, and test plans. Access depends on the API token and project permissions configured for the connection.
- How is authentication configured for this MCP server?
- For local stdio mode, configure `RP_HOST`, `RP_API_TOKEN`, and optionally `RP_PROJECT` in the MCP client environment. For HTTP mode, deploy the server with `MCP_MODE=http` and pass the ReportPortal token as an `Authorization: Bearer` header, with the project key supplied through `X-Project` when needed.
- Which transport should be used for this MCP server?
- Use stdio for local desktop or IDE integrations, especially the documented Docker command using `reportportal/mcp-server`. Use HTTP or SSE only when the MCP server has been separately deployed in HTTP mode and is reachable at `/mcp` or `/api/mcp`.