# CanonicAI Documentation

Corpus in, canonical data out — the corpus-to-dataset engine.

Host: https://docs.canonicai.com
Tier: T-OPEN

## What this host offers
- Product documentation for CanonicAI, the corpus-to-dataset engine: concepts, architecture, data contracts for downstream consumers, and trust/provenance guarantees, plus a generated REST + MCP interface reference.
- Reader-facing docs only; this host does not run extraction or hold corpus content itself — see the REST/MCP reference for the callable surfaces.

## Canonical pages
- [CanonicAI Documentation](https://docs.canonicai.com/) — The corpus-to-dataset engine — corpus in, canonical data out, with provenance.
- [Architecture](https://docs.canonicai.com/architecture) — The pipeline (collector → organizer → referee → outputs), the packages, and the asset registry spine.
- [Concepts](https://docs.canonicai.com/concepts) — The vocabulary of the corpus-to-dataset engine.
- [Data Contracts & Consumers](https://docs.canonicai.com/data-contracts) — CanonicAI's real interface — canonical-output schemas and the federation spine downstream consumers read.
- [Getting Started](https://docs.canonicai.com/getting-started) — Run the pipeline on one book, read its canonical output, and verify the registry.
- [REST API Reference (generated)](https://docs.canonicai.com/reference-apis.generated) — Auto-generated from docs/capability-manifest.json — do not hand-edit.
- [MCP Tools Reference (generated)](https://docs.canonicai.com/reference-mcp.generated) — Auto-generated from docs/capability-manifest.json — do not hand-edit.
- [Trust & Provenance](https://docs.canonicai.com/trust-and-provenance) — Lineage, idempotency, the clean store, and controlled-cost reliability — the producer's guarantees.
- [Trust Center](https://docs.canonicai.com/trust)

## Full content list
- [CanonicAI Documentation](https://docs.canonicai.com/) — The corpus-to-dataset engine — corpus in, canonical data out, with provenance.
- [Architecture](https://docs.canonicai.com/architecture) — The pipeline (collector → organizer → referee → outputs), the packages, and the asset registry spine.
- [Concepts](https://docs.canonicai.com/concepts) — The vocabulary of the corpus-to-dataset engine.
- [Data Contracts & Consumers](https://docs.canonicai.com/data-contracts) — CanonicAI's real interface — canonical-output schemas and the federation spine downstream consumers read.
- [Getting Started](https://docs.canonicai.com/getting-started) — Run the pipeline on one book, read its canonical output, and verify the registry.
- [REST API Reference (generated)](https://docs.canonicai.com/reference-apis.generated) — Auto-generated from docs/capability-manifest.json — do not hand-edit.
- [MCP Tools Reference (generated)](https://docs.canonicai.com/reference-mcp.generated) — Auto-generated from docs/capability-manifest.json — do not hand-edit.
- [Trust Center](https://docs.canonicai.com/trust)
- [Trust & Provenance](https://docs.canonicai.com/trust-and-provenance) — Lineage, idempotency, the clean store, and controlled-cost reliability — the producer's guarantees.

## Machine-readable surfaces
- robots.txt: https://docs.canonicai.com/robots.txt
- sitemap.xml: https://docs.canonicai.com/sitemap.xml
- llms.txt: https://docs.canonicai.com/llms.txt
- mcp.json: https://docs.canonicai.com/.well-known/mcp.json
- openapi.json: https://docs.canonicai.com/openapi.json