> For the complete documentation index, see [llms.txt](https://docs.meganova.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.meganova.ai/libra-os.md).

# Libra OS

> **Note:** Libra OS is MegaNova's enterprise AI operating system — a single-binary platform for running multi-agent workflows, a private knowledge base, and structured research reports over your own documents.

## Overview

Libra OS is what powers MegaNova's vertical products (compliance review, document QA, research reports) and is also available as a deployable platform for enterprise customers.

In a single binary you get:

* **Agents & chat** — ready-made agents plus the ability to create your own, bound to your knowledge
* **Private knowledge base** — upload PDFs, Word docs, Markdown, and more; Libra OS auto-indexes them into a searchable vector store
* **Research reports** — generate multi-section, cited reports from your knowledge plus optional live web research
* **Observational memory** — agents remember what you've told them across sessions, without storing raw chat transcripts
* **Per-user isolation** — every user sees only their own documents and memory by default; admins see the whole tree
* **Single sign-on** — built-in OIDC provider so LibreChat, Chatwoot, or any OIDC client can delegate login to Libra OS

## How to use it

* [**Capabilities**](/libra-os/capabilities.md) — the full map of what Libra OS can do, grouped by task, with the endpoint behind each feature
* [**Use Cases**](/libra-os/use-cases.md) — end-to-end guides for building common products (legal research assistant, doc Q\&A, reports…)
* [**Install**](/libra-os/install.md) — pull the public Docker image and stand up a self-hosted instance in under five minutes
* [**Documents & Knowledge**](/libra-os/documents-and-knowledge.md) — upload files, check indexing, search, organize collections, delete safely
* [**Admin Guide**](/libra-os/admin-guide.md) — system status, agent-to-collection bindings, troubleshooting
* [**Anthropic SDK Quickstart**](/libra-os/anthropic-sdk-quickstart.md) — point the official Anthropic SDK at Libra OS and have your existing code work unchanged (Messages API, Claude Agent SDK, Managed Agents)
* [**Releases**](/libra-os/releases.md) — what's new in each version, migration notes, how to pin to a specific release

## Key concepts in one minute

| Term           | What it means                                                                                                                                   |
| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| **Document**   | A file you upload (PDF, Word, Markdown, etc.)                                                                                                   |
| **Collection** | A named bucket of indexed content. Every user has a private collection; admins can create shared collections.                                   |
| **Chunk**      | A passage of your document after indexing. Agents retrieve chunks to answer questions.                                                          |
| **Agent**      | A configured AI assistant bound to one or more collections.                                                                                     |
| **Report**     | A multi-section markdown document the system generates on a topic, citing your collections as evidence.                                         |
| **Super Nova** | The built-in admin agent. Type `/pause`, `/resume`, `/stats` to control indexing, or ask it plain-language questions about your knowledge base. |

## Next steps

Start with [Documents & Knowledge](/libra-os/documents-and-knowledge.md) — it walks through the most common workflow (upload → search → cite → delete) step by step.