← All resources
Glossary
The vocabulary, defined
A real glossary rather than a keyword page. Where a term is used loosely elsewhere in this category, the entry says what it does not mean — because that is usually the part that matters.
Narraxio terms
- Content Asset
- An individually addressable piece of content with a stable identifier — a headline, a paragraph, a CTA, an FAQ answer, a structured block. Not a page, and not a document: a page is a composition of assets.
- Content Graph
- Content assets plus the typed relationships connecting them to brands, products, facts, audiences, pages and source locations. A graph in the relational sense; a graph database is not required to implement one.
- Source Fact
- An approved business or product truth with an owner, a reviewer, evidence and a validity window, kept separate from the marketing copy derived from it. The thing a claim can be checked against.
- Brand Brain
- The governed context for one brand: voice, vocabulary, forbidden terms, messaging, audiences, approved claims and guardrails — as records a system can read, not as a document a person is supposed to remember.
- Source Binding
- The exact address an asset occupies in source: repository, file, export, object, field. Resolved through the syntax tree so it survives unrelated edits. Without one, safe write-back is not possible.
- Content Provenance
- The separate record of where content originated, who committed it, who claims authorship, what AI touched it, which facts it used, who reviewed it, who approved it and where it was published. Six questions, not one field.
- Content Health
- Continuous analysis of the graph for stale claims, duplication, inconsistent terminology, missing provenance, unsupported claims and content gaps. Findings, with evidence, that a person confirms.
- Content Release
- An immutable numbered snapshot of content that a build can consume, so a published site never depends on a control service being reachable.
- AI Provenance
- The record of which provider and model produced a proposal, under which instruction, from which facts and assets, and what a human changed afterwards.
Industry terms
- Headless CMS
- A content management system that stores structured content and delivers it over an API, leaving presentation to the application. Solves delivery; does not, on its own, solve knowing what content you already have.
- Structured content
- Content stored as typed fields and relationships rather than as formatted documents. A TypeScript object holding a feature's headline and points is structured content, even though nobody involved called it that.
- Content operations
- The practices around producing, governing and maintaining content at scale — inventory, workflow, reuse, localization and measurement. The discipline Narraxio is tooling for.
- Semantic search
- Retrieval by meaning rather than by exact terms, usually through vector embeddings. Excellent for "find copy that says roughly this" and unnecessary for "everything in this repository changed before March".
- RAG
- Retrieval-Augmented Generation: fetching relevant material and giving it to a model before it answers. Narraxio's retrieval is hybrid — structured filters, full text, vectors and graph relationships — because filters alone or vectors alone both miss things.
- MCP
- Model Context Protocol: an open standard for exposing tools and data to AI clients. Relevant here as a possible future surface for the content graph, not as something Narraxio ships today.
- GEO
- Generative Engine Optimization: structuring content so AI answer engines can use it. Mostly the same work as having a coherent content model with clear entities and consistent claims — which is the point of the product anyway.
- Content activation
- Using governed content in channels — campaigns, email, advertising — rather than only publishing it to a website. In Narraxio it is the far end of the same graph: the assets a campaign uses are the assets the website uses.
- Attribution
- Assigning credit for a conversion across the touchpoints that preceded it, according to a chosen model. Produces numbers; does not establish causation, and Narraxio will not describe it as though it does.
Keep reading
The rest of the documentation
Give your content a memory. Give your AI the full picture.
Connect a repository read-only and see what has been sitting inside it — every headline, claim and near-duplicate paragraph, with the file and field each one came from.
Keep Git and your existing stack Read-only to start Narraxio runs on Narraxio