What is M00N Report?

What M00N Report Is

Everything a QA team owns lives in one place.

• The test case library. Manual and automated cases in a folder and suite tree.
• The manual test plan. Executions with assignees, environments, dates and per-step results.
• Automated results. Streaming out of CI while the run is still going.
• The release. Manual and automated evidence rolled into one readiness view.

Cases are the single source of truth for both manual and automated testing. A case can be walked through by a person on Tuesday and covered by a Playwright test on Wednesday without living in two systems.

Where M00N Report Sits in the Category

Test management has moved through four generations.

• Spreadsheets and Confluence pages. A case was a row someone remembered to update.
• Standalone manual-first tools. Built when a test meant a person following numbered steps.
• Jira-native add-ons. The case library moved inside the issue tracker, and inherited its limits.
• AI-native. The same repository a human edits in the UI is fully readable and writable by an autonomous agent, and automated results land in it live rather than being uploaded as a report at the end.

M00N Report is built for the fourth.

The Entity Model at a Glance

Seven entities carry almost everything you will do in M00N Report. They nest inside an organization, which is the tenant boundary.

• Projects - a product, service or repository. A project owns its own case library, folder tree, API keys, tags, integrations and data retention window. Everything below lives inside one.
• Folders and Suites - the shape of the case repository. Folders nest and group; suites are the leaf containers that actually hold cases. Deleting a folder archives it rather than destroying its contents.
• Test Cases - the reusable specification, written once and used everywhere. Rich-text steps with expected results, priority P0 to P4, status, tags, attachments, custom fields and dependencies. Cases can be typed by hand, imported from CSV, or authored by an AI assistant over MCP, and any case can be linked to the automated test that covers it.
• Test Collections - named, reusable sets of cases assembled across folders and suites: a smoke pack, a regression pack, a payments pack. Build the collection once and drop it into as many executions as you need.
• Manual Test Executions - the planned manual run. An execution carries assignees, an environment, start and target dates, and per-step pass, fail, blocked or skipped results. Each case is snapshotted as it enters the execution, so the record of what was actually tested does not shift when someone edits the case next month.
• Releases - the milestone. Manual executions and automated launches both attach to a release, and their combined pass rate becomes the readiness signal for shipping. Releases carry version numbers, target dates and sprint associations. Deleting a release archives it.
• Launches - automated runs streamed in from CI. A launch holds suites and individual test results with retries, durations, error output and attachments such as screenshots, video, traces and logs, and it updates live while the run is still going.

How Work Gets Into M00N Report

There are three doors in. Two of them carry automated results from your test runners. The third lets an AI assistant author and manage the test assets themselves.

Official Reporters

Five reporters are built and maintained by us, published on the registry your language already uses: Playwright and Jest on npm, pytest on PyPI, NUnit on NuGet, and JUnit 5 on Maven Central. Add the package, point it at your server URL with a project API key, and run your tests exactly as you did before.

Results stream while the run is still in progress rather than being uploaded after it finishes, so a failure shows up in the UI the moment it happens. Reporters are written to fail silently: if M00N Report is unreachable, your test run still completes normally.

The OpenAPI Ingest Contract

If your language or runner has no official reporter, the ingest API is fully specified as an OpenAPI 3.0.3 document published at /api/ingest/v2/openapi.json. Generate a client from it, or hand the spec to a coding assistant and have it write the reporter for you. It is the same contract the five official reporters speak, so nothing about the resulting data is second class.

The published specification covers result ingestion. Managing cases, executions and releases is done through the UI and through MCP.

MCP for AI Assistants

M00N Report ships a Model Context Protocol server with 55 tools: 25 read-only, 31 that change state, and 9 flagged as destructive so a client can prompt before running them. It also ships prompts that run real queries against your data and resources for browsing project structure.

The state-changing count is the number that matters. An assistant connected to M00N Report does not just read a dashboard, it works: it authors test cases from Markdown into the editor's own format, creates the folder and suite hierarchy and tags they need, assembles collections, creates and populates manual executions, links automated tests to the cases they cover (including before that test has ever run), pulls coverage gaps in both directions, runs the composite health check, and downloads failure traces and screenshots to reason about them.

Because the agent works through the same service layer as the UI, agent-authored tests land inside the same roles, permissions and audit log as everything else. That is the point: AI writes a lot of tests, and this is where they get managed.

There are two ways to connect. The hosted remote connector uses OAuth with a browser consent screen and needs no key file or local process; its redirect allowlist covers claude.ai and claude.com only. Every other MCP client (and any self-hosted instance) connects through the stdio npm package authenticated with an MCP key.

Who M00N Report Is For

Where to Go Next

Read these in order and you will have the whole model.

• Core Concepts - organizations, projects, API keys, launches, cases, executions, releases, tags and roles, defined precisely and in one place. Start here.
• Test Cases, Manual Test Executions and Releases - the test management side, page by page: how the folder and suite tree behaves, how to plan and run an execution, and how a release turns manual and automated evidence into a readiness call.
• Reporters - pick your runner and get automated results streaming: Playwright, Jest, JUnit 5, pytest or NUnit. If yours is not listed, the Integration API guide walks the OpenAPI ingest contract.
• MCP Quick Start - connect an AI assistant and put the 55 tools to work.
• If you would rather see it working before you read about it, Getting Started has your first run streaming in about five minutes.

Common Questions

Is M00N Report a test management system or a reporting dashboard?

A test management system. It owns the test case repository, the folder and suite tree, reusable collections, manual executions with per-step results, and releases. Real-time automated results are one part of it, not the definition of it.

Do I need CI or automation to use it?

No. Cases, collections, executions and releases work with zero automated tests connected. Add reporters later and the same cases start showing automated coverage.

Can an AI assistant really create test cases, or just read them?

Create. Of the 50+ MCP tools, 31 change state, including creating cases, folders, suites, collections, executions and releases. Nine are flagged destructive so your client can confirm first.

Which test runners have official reporters?

Five: Playwright, Jest, JUnit 5, pytest and NUnit, published on npm, Maven Central, PyPI and NuGet. Anything else integrates through the OpenAPI 3.0.3 ingest contract.

Can I run it on my own infrastructure?

Yes. Self-hosted deployment is supported with Docker Compose or Kubernetes, including SSO, roles and the MCP server.