3 snapshots recorded · history starts from the first webhook event after deploy
3Latest version
---
slug: what-is-usable
−section: Getting Started
+section: Understanding Usable
order: 1
status: published
description: "What Usable is, how its parts fit together, and when it is the right tool."
---
··· 58 unchanged lines ···
## Continue with the core concepts
Start with [Workspaces](/usable/core-concepts/workspaces) to understand where knowledge lives, then read [Memory fragments](/usable/core-concepts/memory-fragments) to learn how individual pieces of knowledge are structured.
2Version 2
---
slug: what-is-usable
section: Getting Started
order: 1
status: published
−description: "What Usable is, what it is not, and when it is the right tool."
+description: "What Usable is, how its parts fit together, and when it is the right tool."
---
1Version 1
Oldest recorded snapshot — no earlier version to compare.
−Usable is a **knowledge platform** built around workspaces of memory fragments — small, structured pieces of information that people and AI agents can create, search, and reuse.
+Usable is a knowledge platform for information that needs to remain useful to both people and AI.
−It sits between a wiki, a notes app, and an AI context layer. You store what your team knows in fragments; Usable Chat, MCP tools, and the API can retrieve and build on that knowledge instead of starting from a blank prompt every time.
+Instead of keeping important context in disconnected chats, private notes, and long documents, Usable stores knowledge as focused **memory fragments** inside **workspaces**. People can read and maintain those fragments, while Usable Chat and connected AI tools can search them and bring the relevant material into a conversation.
+The result is a shared knowledge base that can support documentation, product context, operational guidance, research, reusable AI skills, and other information your team expects to use again.
+
+## The basic model
+
+A workspace defines a boundary around a body of knowledge. Within that workspace, each fragment covers a useful topic: a concept, a procedure, a known solution, a template, or another kind of information.
+
+Fragment types communicate what kind of information a fragment contains. Tags, summaries, collections, and metadata provide additional ways to describe and organize it. Search then uses those signals together with the fragment content to help people and agents find the right source.
+
+This structure is deliberately flexible. A small team might use one workspace for all internal knowledge. A larger organization might separate public documentation, engineering decisions, support material, and customer-specific information into different workspaces.
+
## What Usable is
−**A workspace for durable knowledge.** Each workspace holds fragments: docs, recipes, solutions, skills, and custom types you define. Fragments support Markdown, frontmatter, tags, summaries, and attachments so content stays findable by both humans and models.
+### A place for durable knowledge
−**A retrieval layer for AI.** Semantic search and agentic tools (including MCP) let Chat and external agents pull the right fragments into context. The point is grounded answers from *your* material, not generic model guesses.
+Usable is designed for information that should survive beyond the conversation or task that produced it. A troubleshooting discovery can become a Solution fragment. A repeatable process can become a Recipe. Product behavior can be explained in a Knowledge fragment.
−**A collaboration surface.** Workspaces have members and roles. You can keep knowledge private, share it with a team, or publish it more widely. Collections help you organize related fragments without forcing a rigid folder tree.
+Good fragments are maintained sources, not snapshots of a chat transcript. They give future readers enough context to understand and apply what was learned.
−**An integration platform.** REST APIs, API tokens, webhooks, embeddable Chat, and MCP mean Usable can sit behind docs sites, bots, IDEs, and internal tools — not only the Usable UI.
+### A context layer for AI
+Usable can provide grounded context to Usable Chat and external AI clients through MCP and APIs. An agent can discover relevant fragments, read their full content, and—with appropriate permission—create or improve knowledge as work progresses.
+
+This does not make every answer automatically correct. Retrieval quality still depends on clear titles, focused content, useful summaries, and sensible organization. Usable gives AI a durable source to consult; people remain responsible for the quality of that source.
+
+### A shared workspace for people and tools
+
+Workspaces can be private to invited members or made public for broader discovery. Collections group related fragments, while fragment types and tags make their purpose easier to understand. Files can be attached where supporting material is useful.
+
+The same knowledge can be used from the dashboard, Usable Chat, MCP-enabled clients, or an integration built with the REST API.
+
## What Usable is not
−**Not a traditional CMS for marketing pages.** You *can* publish documentation from fragments (this site does), but Usable is optimized for knowledge units and AI retrieval, not pixel-perfect page builders or blog themes.
+Usable is not a replacement for every system your team already uses.
−**Not a replacement for your source-of-truth code repo.** Fragment knowledge should complement git — decisions, runbooks, product context, and “why we did it this way” — not replace versioned application code.
+It is **not a source-code repository**. Keep application code and version-controlled configuration in Git. Use Usable for the explanations, decisions, runbooks, and product context that make that code easier to understand and operate.
−**Not a generic chatbot with no memory.** Chat without a workspace full of good fragments is underpowered. Usable’s value compounds when you write clear, maintained fragments and keep them tagged and typed.
+It is **not merely a chatbot**. Chat is one way to interact with workspace knowledge, but the fragments remain useful independently as readable and searchable documentation.
−**Not only for AI engineers.** Anyone who writes product notes, support playbooks, onboarding guides, or research can use it. AI features amplify well-written knowledge; they do not replace authorship.
+It is **not a pixel-perfect website builder**. Fragments can power public documentation and other publishing workflows, but Usable’s core concern is structured, reusable knowledge rather than visual page design.
−## When Usable fits
+It is also **not automatic knowledge quality**. Importing raw logs, duplicating documents, or generating vague summaries creates noise. Usable works best when teams treat fragments as maintained material with clear ownership and purpose.
−Use Usable when you want a shared place for team knowledge that both people and agents can search, cite, and update — documentation, operational memory, product context, and reusable skills in one system.
+## When Usable is a good fit
−If you only need a static brochure site with no AI retrieval, a simpler docs tool may be enough. If you need agent-ready, workspace-scoped knowledge with Chat, MCP, and APIs, Usable is built for that.
+Usable is useful when knowledge needs to be:
+
+- shared across people, applications, and AI agents;
+- searchable by meaning as well as exact metadata;
+- separated into clear workspace boundaries;
+- improved over time instead of repeatedly rediscovered;
+- available through a UI, Chat, MCP, or APIs.
+
+If all you need is a static brochure site or short-lived notes for one person, a simpler tool may be enough. If you need an evolving knowledge base that humans can read and AI can reliably retrieve, Usable is built for that role.
+
+## Continue with the core concepts
+
+Start with [Workspaces](/usable/core-concepts/workspaces) to understand where knowledge lives, then read [Memory fragments](/usable/core-concepts/memory-fragments) to learn how individual pieces of knowledge are structured.