# Tagvico AI V3 documentation

> Official, versioned documentation for Tagvico AI, a self-hosted AI filing companion for Paperless-ngx.

Use these instructions for v3. Do not mix commands or configuration with another major version. Tagvico reads Paperless OCR text and metadata; hosted model providers receive the content needed for classification. Start with Review first mode and pin an immutable container image tag.

## V3 guide

- [Overview](/docs/index.html.md): What Tagvico AI does, its operating modes, and where to start.
- [Install Tagvico v3](/docs/installation.md): Install the current Tagvico release with Docker Compose or docker run and complete guided setup.
- [Upgrade to or within v3](/docs/upgrading.md): Back up, upgrade, validate, and roll back a Tagvico installation safely.
- [Remove Tagvico AI](/docs/removing.md): Remove the container, optionally delete local data, and revoke credentials.
- [Feature showcase](/docs/features.md): Review filing, Action Center, Companion approvals, household access, OCR, and model discovery.
- [AI tagging and metadata for Paperless-ngx](/docs/paperless-ai-tagging.md): Compare Tagvico with built-in Paperless matching and evaluate local or hosted AI metadata safely.
- [Model provider overview](/docs/providers.md): Compare local, hosted, direct, compatible, and subscription-backed model providers.
- [Privacy and security](/docs/privacy.md): Understand data flow, secret storage, deployment boundaries, and screenshot policy.
- [Troubleshooting](/docs/troubleshooting.md): Diagnose setup, connectivity, provider, processing, and upgrade problems.
- [Release notes](/docs/release-notes.md): Read the release notes guide.

## Optional

- [Complete V3 documentation](/docs/llms-full.txt): All guide pages combined into one Markdown document.
- [Source repository](https://github.com/arturict/tagvico-ai): Application source, releases, and issue tracker.
- [Product website](https://tagvico.arturf.ch/): Tagvico AI landing page.