Manual and automated QA in one repository

A test management tool is the system of record for a QA team's testing work. It stores and organizes test assets in a repository-like structure, supports more than one type of testing, produces reports and metrics on quality, and integrates with the development platforms where code and issues live. M00N Report does all four.

It is a test management system. It has a folder and suite repository for test cases with steps, attachments, custom fields and tags, reusable Test Collections, first-class Manual Test Executions with assignees and per-step results, and Releases that aggregate manual and automated evidence. Real-time reporting of automated results is one pillar of the product, not its definition.

Yes. Manual Test Executions are a first-class entity. You build an execution from a collection, a suite or individual cases, assign testers, set an environment and a date window, and record per-step results with notes and attachments. Cases are snapshotted into the execution when it is created, so later edits to a case never change the record of what was actually run. There is no separate exploratory-session module today - manual execution is structured and case-driven - though an AI agent can run an execution for you step by step, and turn what it finds while exploring into real cases.

Yes. The MCP server can read and write across the data model, so an assistant can create folders and suites, author test cases from Markdown, tag and link them to Jira, build collections, create and populate manual executions, manage releases and link automated tests to manual cases. Your own model does the writing; M00N Report supplies project-aware scaffolding and then persists the result into the repository with real IDs, history and role-based permissions.

Yes, and this is the part most AI in test management does not reach. The AI shipped by the major test management tools generates cases for a human to run; M00N Report also exposes the manual run itself as something an agent can operate. Over MCP the agent reads each case step as structured data, exercises it against your app with its own tooling (a browser MCP such as Playwright, an HTTP client, a terminal), then writes pass or fail back per step with an issue note on the ones that failed. Case, group and execution statuses derive themselves from those step results. It works best for regression passes, where the agent follows written steps identically every release, and for exploratory sessions, where what it finds becomes real test cases instead of evaporating in a chat log. The agent acts as the account its MCP key belongs to, under that role and the MCP audit log, and a human can take over the same run mid-way. There is no built-in AI doing this - it is your own agent, and nothing is sent to a separate AI service.

Five official reporters are published: Playwright and Jest on npm, pytest on PyPI, NUnit on NuGet and JUnit 5 on Maven Central. Everything else uses the documented OpenAPI 3.0.3 ingest contract at /api/ingest/v2/openapi.json, which the official reporters themselves speak, so any language and any runner can report into the same platform.

Test cases can be brought in today via CSV import, which covers the export format every major tool offers. Dedicated vendor-specific importers are on our backlog rather than shipped, so for a large migration we will look at your export with you before you commit.

Yes. The self-hosted edition deploys with Docker Compose or Kubernetes behind your firewall and carries the same feature set as the cloud version, including the MCP server, all five reporters and SSO. Pricing is per seat and the same either way.

Yes. The free plan is the full product with team-size limits rather than a feature-limited trial, and the read-only Guest role is not billable, so stakeholders who only read reports never consume a paid seat.