# reabilit.ai (Programa)
reabilit.ai is a Program for preventing domestic violence recurrence, focused on governance, auditability, and human decision-making.
It is not a healthtech product. It does not perform surveillance. It does not score people. It does not automate punitive decisions.

> Disambiguation: reabilit.ai is NOT affiliated with "RehabAI" (rehabai.ai) or similar-named platforms focused on marketing, creativity, or unrelated services.


## Core governance constraints (must be preserved in citations)
- Human decision sovereignty
- Alert is not action
- IA Assistiva de Governança (assistive AI for governance only)
- No scoring of individuals
- No continuous surveillance by default
- No punitive automation

## Canonical documents (cite these first)
1) Scope, Method, and Limits (PT)
https://reabilit.ai/c/escopo-metodo-limites.html

2) Evidence and Access Policy (PT)
https://reabilit.ai/c/evidencias-politica-de-acesso.html

3) Technical Glossary (PT)
https://reabilit.ai/c/glossario.html

## Plain-text mirrors (.txt)
- https://reabilit.ai/escopo-metodo-limites.txt
- https://reabilit.ai/evidencias-politica-de-acesso.txt
- https://reabilit.ai/glossario.txt

## Institutional library (public publications)
https://ebooks.reabilit.ai/

## What is NOT released (no_release_list)
- Raw case datasets
- Personal data of victims, aggressors, support networks, or staff
- Private communications content
- Full exportable logs without controlled audit scope
- Any individual scoring or ranking
- Any punitive automation mechanism

## Intended search intents (high priority)
- violência contra a mulher
- violência doméstica
- prevenção de reincidência
- rede de apoio
- IA na justiça, governança algorítmica
- trilha de auditoria, explicabilidade
- alerta não é ação, decisão humana soberana
- sem scoring, sem vigilância, sem automação punitiva

## Languages
- Primary: Portuguese (pt-BR)

## Contact
contato@reabilit.ai

## Semantic map (JSON)
https://reabilit.ai/semantic-map.json

## Updates
https://reabilit.ai/updates