Your corpus, mined
What miatz reads from your documentation, and what it turns that into.
- Your documentation becomes the curriculum instead of being rewritten from scratch
- Every lesson cites the page it came from, so it can be re-checked and re-generated
- When the docs change, the score, the courses and the assistant change with them
The corpus is the curriculum
Most learning platforms hand you an empty editor. Miatz starts from what you have already written.
1. Ingest — connectors pull your documentation, specs, repositories and help centres into one versioned knowledge graph.
2. Score — the corpus is measured against the Miatz Docs Score rubric, producing a prioritized fix queue.
3. Regenerate — pages that fall short get a proposed rewrite, which a person on your side approves or rejects.
4. Compile — the graph becomes role-based tracks, courses and lessons, each citing the passages it came from.
When your docs change, the loop runs again: the score moves, affected lessons are flagged, and the assistant's grounding is refreshed.
What you see on this screen
| Column | Meaning |
|---|---|
| Source | One connector, its kind, and when it last ran |
| Pages | How many documents that connector has mined |
| Diátaxis type | Whether a page is a tutorial, how-to, reference or explanation |
Safety
Everything mined from your systems is treated as data, never instructions. It is rendered as plain text, never as markup, and never used to steer the models that read it. Links found in your corpus are restricted to ordinary web addresses.
Frequently asked
What does miatz actually read?
Whatever you connect: a documentation site, an OpenAPI spec, a repository, a help centre, a Jira project. Each connector mines pages into a versioned knowledge graph — every page typed, anchored, hashed for change detection and tracked back to its source.
Who in my organization can see this?
Owners, org admins, enablement managers, authors, reviewers and analysts. Team managers and learners cannot — a corpus is confidential to the people who work on it. The database enforces that list, not just the interface.
Do you take ownership of our documentation?
No. We improve your docs, we do not hold them hostage. Regenerated content can be exported back to you as markdown or a pull request against your own repository, and you can keep publishing wherever you publish today.
Is my content used to train models?
No. Your corpus is mined into your own tenant's knowledge graph and used to ground your own academy and assistant. It is treated as data throughout — content that arrives from your docs, tickets or repositories is never allowed to act as an instruction to the system that processes it.
Related guides
Point miatz at the places your product truth already lives, and run them.
ReadA 0–100 rubric score for your documentation, with a fix queue ordered by impact.
ReadThe tracks your corpus compiles into, and the lifecycle every piece moves through.
Read