Skip to content
← All resources

Documentation

Concepts

What the objects in Narraxio are, how they relate, and which of them exist today. This is the conceptual reference; the developer and API pages cover the surfaces.

The core objects

Six objects carry most of the model. Everything else is a view over them.

Content Asset
The unit. A headline, a paragraph, a CTA, an FAQ answer, a proof point, a meta description or a structured block — with a stable identifier, a type, a language, a status and a source binding. The identifier survives rewording, which is what makes history, reuse and eventually performance data attachable.
Source Binding
The exact address an asset came from: repository, file, export, object, field. Resolved through the syntax tree rather than by line number, and re-verified against the latest revision before any write. Without it, editing source is guesswork.
Source Fact
An approved business or product truth with an owner, a reviewer, evidence and a validity window — kept deliberately separate from the copy derived from it. Facts are what make a claim checkable and what make a product change into a content work queue.
Brand
A voice, a vocabulary, a set of audiences and a permission boundary. An organization can hold several; isolation is the default and sharing is an explicit grant.
Relationship
A typed edge between objects: belongs to brand, describes product, derived from fact, appears on page, in section, variant of, translated from, reuses, supersedes, targets audience. Typed rather than tagged, because the consequences differ.
Version / Event
An append-only record per asset. Imported, edited, AI-suggested, reviewed, approved, published — each a separate event with an actor and a timestamp. Versions are added, never rewritten.

Content asset fields

The shape an asset carries. Typed and nested values are preserved rather than flattened into text.

  • content_asset_id, organization_id, brand_id, optional product_id
  • type, language, value or structured value
  • status, intent, funnel_stage, audience_ids
  • source_binding_id
  • created_at, first_seen_at, last_changed_at, published_at
  • provenance summary and current_version_id

Workflow states

Seven states, and the reason AI Suggested is one of them separately from Draft: reviewers read the two differently, and collapsing them removes the signal they most need.

  • Imported — came from a source, untouched
  • Draft — a person is working on it
  • AI Suggested — a model proposed it, nobody has accepted it
  • In Review — submitted against the policy for its content type
  • Approved — cleared, not yet published
  • Published — live at its publication locations
  • Archived — retained with its history, no longer in use

The modules, and what each one owns

Eight modules over one graph. The table is the quickest way to see which one a given question belongs to.

ModuleOwnsThe question it answers
Content GraphAssets, relationships, searchWhat do we have, and where does it appear?
Models & SourcesParsers, content types, bindingsWhat shape is it in, and where exactly did it come from?
Brand BrainVoice, facts, messaging, audiences, guardrailsWhat are we allowed to say, and on what evidence?
AI StudioAsk, brainstorm, create, rewrite, repurpose, translateWhat should this say instead?
Provenance & GovernanceAuthorship, AI records, versions, approvalWho wrote it, what touched it, who signed it off?
Content HealthDrift, duplication, unsupported claims, gapsWhat is wrong with what we have?
Publish & SyncPatches, pull requests, releases, API, webhooksHow does an approved change reach production?
Performance & ActivationExposure, conversions, revenue, campaignsDid any of it work?

Where to go next

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