Skip to main content

Documentation Index

Fetch the complete documentation index at: https://kasava.dev/docs/llms.txt

Use this file to discover all available pages before exploring further.

A workspace is a bounded context container for a specific product initiative — “Handbook Alpha,” “Mobile Checkout Redesign,” “Search Performance.” It’s where you collect evidence, form and test hypotheses, record decisions, and turn all of that into specs that are ready for engineering. Workspaces sit inside a product. One product has many workspaces, one per initiative or problem space.

Why workspaces exist

There’s a gap in most product toolchains between “what are customers saying?” (Dovetail, Grain, user interviews) and “what are we building?” (Linear, Jira, GitHub). The messy middle — where evidence becomes conviction, conviction becomes decisions, and decisions become specs — usually lives in scattered Notion docs and Slack threads. Workspaces own that middle. They give you a single bounded space for the upstream product work that happens before a plan crystallizes.

Anatomy of a workspace

Workspace homepage
A workspace contains:
  • Context — an AI-maintained summary of what this initiative is about, updated as you add material
  • Sources — documents, URLs, notes, uploads, and conversation transcripts that serve as evidence
  • Hypotheses — explicit beliefs with links back to the evidence that supports or contradicts them
  • Decisions — a decision journal with recorded rationale, alternatives, and outcomes
  • Plans — when the thinking is solid, formalized plans link out to documents and work items
  • Workspace-scoped chat — every message to the Kasava Agent inside a workspace gets the workspace’s context auto-injected

How context injection works

Everything scoped to the workspace — chat messages, plan generation, document drafting — receives the workspace’s instructions, sources, hypotheses, and decisions as context. You don’t paste the same background into every prompt. The Agent already has it. This is how workspaces compound: the more evidence and decisions you accumulate, the better every subsequent operation gets.

Workspace vs. Product vs. Organization

ScopeWhat it isExample
OrganizationYour tenant — billing, members, access controlAcme Inc.
ProductA software product you’re buildingAcme Web App
WorkspaceA specific initiative inside a productMobile Checkout Redesign
You can have many workspaces per product, and you can move between them without losing context — each keeps its own sources, decisions, and chat history.

When to create a workspace

Create a workspace when you’re starting to learn, believe, or decide about something:
  • A new initiative you’re scoping out
  • A known problem area you’re digging into
  • An experiment you’re running and tracking results from
  • A competitor response you’re planning
Workspaces are team-visible by default. The evidence, hypotheses, and decisions persist so future teammates can see how a decision was made — not just what was decided.

The Product Graph

The code- and commit-grounded knowledge model workspaces plug into

Plans & Work Items

What happens when workspace thinking is ready for execution

The Kasava Agent

The chat agent, with workspace context auto-injected

Products

The container that holds your workspaces