RAG memory for the Visualiza ecosystem
search project knowledge without training a model
Summary and current operating status.
Overview
Visualiza Brain is a planned RAG layer for approved ecosystem knowledge.
Scoped retrieval preview
Visualiza Brain organizes source categories, previews dry-run search and shows privacy boundaries before any live index, embedding job or answer service is connected.
Brain Search
A safe interface preview for retrieval over scoped Visualiza knowledge.
No indexed sources yet.
After indexing, Visualiza Brain will show matching sources and answer with source references.
Dry run only. No embeddings, vector DB or LLM call yet.
Source Registry
Static source categories only; no indexed records or private source content are shown.
Source Registry controls what Visualiza Brain may search.
Private sources are never exposed globally.
No embeddings or vector search are active yet.
StyleCards
Future retrieval labels for approved style direction.
Projects
Project-level context once scoped sources are connected.
Prompts
Private prompt text stays unavailable until explicit scoped indexing exists.
Renders
Render metadata only; no private assets indexed here.
Ratings
Rating signal categories for future retrieval filters.
Training Pool
Consent-aware references stay separate from training use.
Label
Public label context after approval boundaries are clear.
Operator Reports
Operational report summaries require admin-only search boundaries.
Agents
Agent notes and handoff metadata for future lookup.
Business Docs
Business documents stay blocked until private access rules are defined.
RAG Flow
The intended retrieval path is visible without connecting a live index.
Ask
A user asks a Visualiza ecosystem question.
Retrieve
Scoped metadata would find relevant source context.
Read Context
The answer layer would inspect only approved snippets.
Answer with Sources
Future responses should cite the records used.
Privacy
The page states boundaries clearly while the Brain remains a UI-only dry run.
Retrieval reads scoped context at answer time; it does not train or fine-tune a Visualiza model.
Personal project context stays inside its approved visibility scope and is not exposed globally.
Operator reports, agent notes and business docs require admin-only search boundaries before indexing.
Retrieval visibility does not grant permission to use an asset for model training.
Public Label visibility only follows explicit publication review and approval.
Environment files, credentials and private runtime secrets stay outside Visualiza Brain sources.