Skip to content
AI Knowledge Governance

AI audit trail & knowledge base versioning for RAG

A knowledge governance platform for LLM applications. Full AI audit trail, schema and content versioning, RBAC, and RAG observability — built into your knowledge base API.

Get StartedLearn More
Schema Versioning
Knowledge base versioning for every schema change. Compare versions. Safe migrations.
Data Versioning
AI audit trail. Instant rollbacks. Know who changed what and when.
Localization
Multi-language content API. Locale-aware search. Fallback chains.
Access Control
Knowledge base access control. API key scoping. Folder-level RBAC.
Migration-Safe
Non-breaking updates. Forward-compatible. Zero downtime.
Dual APIs
Management API for writes. Flux API for reads. Full control.
Schema Versioning

Schema versioning API — track every change

Your schema evolves as your product evolves. The schema versioning API tracks every change and makes it reversible — a core part of LLM knowledge base management.

Schema History
articles.schema.json
{
  "fields": [
    {
      "key": "title",
      "type": "string"
    },
    {
      "key": "content",
      "type": "text",
      "meta": { "vectorizable": true }
    },
    {
      "key": "category",
      "type": "string"
    },
    {
      "key": "slug",
      "type": "string"
    }
  ]
}

Version Comparison

See exactly what changed between versions.

Safe Rollback

Revert schema to any previous version.

Non-Destructive

Existing data preserved. Update when ready.

Data Versioning

AI audit trail — every edit, every version

Each save creates an immutable revision with a full AI audit trail. Content versioning API preserves every change — restore any version instantly for RAG pipelines or rollback after errors.

1

Edit content

Make changes in dashboard or via API

2

New revision created

Auto-incremented number, timestamp, author tracked

3

Previous version preserved

Nothing deleted. Restore anytime with one click.

Revisionsfaq/refund-policy
#3
SarahPublished

Jan 10, 14:32

#2
Mike

Jan 10, 11:05

Restore
#1
Sarah

Dec 15, 09:22

Restore
Localization

Multi‑language content API with fallback chains

Store content in multiple languages via a built-in multi-language content API. Automatic fallback to default locale. Single embedding handles queries in any language.

Content Structure
ENdefaultRefund Policy
DERückerstattung
FRfallback → EN
Missing translations fallback to default locale
GET ?return_locales=en,de,fr&fallback_locales=true
{
  "data": {
    "title": {
      "en": "Refund Policy",
      "de": "Rückerstattung",
      "fr": "Refund Policy"  // fallback
    }
  }
}

Single Embedding

One vector from default locale. Fewer tokens, same quality.

Automatic Fallback

Missing translations filled from default locale automatically.

Cross-Locale Search

Query in one language, find content in another.

Access Control

Knowledge base access control with RBAC

Granular knowledge base access control. RBAC for API keys — define who can read, write, or delete at the folder level. Scoped API keys for external access.

Roles & Permissions
RoleCollectionsActions
AdminAllRead, Write, Delete, Publish
Editorarticles, faqRead, Write
ViewerarticlesRead

Scoped API Keys

Create API keys with specific permissions. Assign to roles that control access to folders and actions.

ReadWriteDelete

Environment Isolation

Separate dev, staging, and production. Same schema, isolated data and permissions.

devstagingproduction
Migration-Safe

Data versioning for AI — iterate without fear

Schema updates are instant and non-destructive. Data versioning for AI means your content, embeddings, and structure stay safe while you evolve — including embedding versioning when you change vectorizable fields.

Instant Updates

Publish schema changes immediately. No downtime, no deployment pipelines.

Data Never Lost

Removed fields are hidden, not deleted. Recover data anytime if needed.

Rollback Anytime

Made a mistake? Revert to any previous schema version with one click.

Independent Layers

Schema and data versioned separately. Update structure without touching content.

Dual APIs

Management & Flux APIs

Data governance for LLM starts with API separation. Management API for full CRUD operations with RAG observability. Flux API for read-only access with vector search. Choose the right tool for each use case.

Management API

Full Control

Create, update, delete resources. Manage schemas and revisions.

  • Create/Update/Delete
  • Schema management
  • Revision history
  • Role-based access
Flux API

Read-Only Access

Fast, cacheable reads. Vector search. Optimized for agents.

  • Hybrid search (vector + text)
  • Advanced filters & joins
  • Localization support
  • Public or authenticated

Use Management API for content operations (CMS sync, migrations, admin dashboards).
Use Flux API for runtime access (RAG, agents, public websites).

AI audit trail & knowledge governance
for production RAG.

Get Started See RAG Example

Explore more

Hybrid Search API

Semantic search, full-text, and pre-filter vector search in one query.

Learn more →

Knowledge Base API

Schema-first, auto-generated REST API with built-in search. Instant backend for AI.

Learn more →

Build RAG Apps

End-to-end guide to building production RAG applications.

Learn more →

AI Agent Memory

Persistent knowledge base for AI agents. Read-write API with LangChain integration.

Learn more →