Managing Knowledge

#Purpose

Every organization already has a knowledge base. It is just usually scattered — the current procedure in one person's drive, last quarter's version in a shared folder, and the answer everybody actually uses somewhere in a chat thread.

Knowledge is Opal's centralized knowledge base: a single, organized home for the documents, files, and shared context your organization relies on, built as a familiar system of files and folders. What makes it different from a document store is who it serves — it is a source of truth that people and agents work from, so the same reviewed content answers a colleague's question and grounds an agent's work overnight.

This subsection covers that system end to end: getting content in, organizing it, turning raw uploads into content that can actually be read and acted on, keeping it accurate through a governed lifecycle, and putting it to work with agents.

It is written for anyone who captures, organizes, and maintains the information their work depends on — knowledge workers, subject-matter experts, and operations teams — as well as people setting up agents to work from that same content. No technical background is assumed.

This page is the map. It introduces each part of the knowledge system, says which page covers it, and tells you where to start.

#In This Section

• Knowledge Overview — A conceptual introduction to the knowledge base: what the pieces are, how they fit together, and the four moves that take content from upload to grounded agent work. Knowledge Overview
• Files & Folders — Adding content by uploading it or authoring it in Opal, organizing it in the centralized file-and-folder structure, and keeping it tidy with rename, move, archive, and sharing. Files & Folders
• Libraries — Standalone collections that bring related files and folders together for a particular purpose, without rearranging the folder structure they already live in. Libraries
• File Processing — How Opal detects a file's type and offers processing suited to it — OCR digitization to markdown, transcription with speaker diarization, annotation, and extraction of embedded artifacts. File Processing
• Viewing & Editing Files — Reading and editing files in an interface built for each format, with version history, an activity log, and a draft-to-published workflow with optional review. Viewing & Editing Files
• Knowledge for Agents — How content added to a space becomes the grounding context agents retrieve in threads, tasks, and flows, how agents contribute documents back, and how each file is delivered in a form the agent's model can use. Knowledge for Agents

#Key Concepts

Content lives in folders; libraries pull it together. The centralized, file-system-like structure of files and folders answers where does this content live? — every file sits in one place in it. A library answers a different question: what belongs together for this piece of work? Libraries gather related files and folders into a standalone collection alongside the structure, rather than instead of it. Most teams use both.

Processing is what turns a file into knowledge. A scanned contract is a picture of words; an hour-long workshop recording is an audio file. When content is added, Opal detects its type — document, image, audio, or video — and offers the processing that fits, so the content can be read, searched, and acted on rather than left locked in its original format.

Publishing is what makes an edit official. Files are edited in place, in a viewer and editor suited to their format, and changes are made in a draft and published as a new version. Version history keeps every published version, any of them can be restored, an activity log records what changed and who changed it, and organizations that turn the publishing workflow on require review and approval before a version goes live.

Agents reach knowledge through spaces. Add files and folders to a space, and the agents working there retrieve that content as grounding context in threads, tasks, and flows — and can create and edit documents themselves, under the same review and publishing controls. Opal delivers each retrieved file in a form the agent's model can use: natively where the model supports that input modality, and as a text equivalent, such as a transcription, where it does not.

Knowledge is not the same as skills. Skills package reusable instructions that shape how an agent performs a kind of work. Knowledge is the reference content an agent draws on to know about your business. Skills supply the method; knowledge supplies the material.

Short definitions of every term used here are in one place: Glossary.

#Where to Start

Start with Knowledge Overview. It sets out the whole system in one read — the pieces, how they connect, and the path content takes from upload to grounded agent work. Everything else in this subsection is detail underneath it.

From there, go where your task takes you:

• To get content in and organized — Files & Folders, then Libraries when related material needs to sit together without being moved.
• To make scans, recordings, and other raw uploads usable — File Processing.
• To edit content and manage its lifecycle — Viewing & Editing Files.
• To put knowledge to work with agents — Knowledge for Agents.

Reading the pages in order works well if the knowledge base is new to you; each page also stands on its own if you have arrived with a specific job to do.

If your next question is where the work that uses this knowledge happens, read Getting Work Done, which covers spaces, threads, tasks, and flows.