Key Concepts

Understanding how Kommit's core pieces fit together will help you get the most out of the platform.

Projects

A project is the top-level container for a product idea. Everything in Kommit lives inside a project: the canvas, all nodes, AI conversations, and generated PRDs.

Each project belongs to a workspace (your team's account). Projects are private to your workspace by default.

Canvas

The canvas is an infinite, zoomable workspace where you organise your product spec visually. Nodes sit on the canvas and edges connect them. Think of it as a living diagram of your product's requirements.

The canvas auto-saves every 30 seconds. You can also access snapshot history to restore earlier versions.

See Canvas overview for details.

Nodes

Nodes are the core unit of a spec. Each node represents one aspect of your product and holds both a conversation and structured data. Kommit has six node types:

Project

The root node. Captures the product's name, description, and high-level goals. Every project starts with one Project node.

Audience

Documents who the product is for. Captures personas, company size, job roles, pain points, and success metrics.

Competitors

Maps the competitive landscape. Captures competitor names, their strengths, their weaknesses, and the gaps your product fills.

Tech Stack

Records the planned or existing technical architecture. Captures frontend framework, backend runtime, database, infrastructure, and third-party services.

Features

Describes individual product capabilities. Captures feature name, description, user stories, priority (P0–P3), acceptance criteria, and complexity.

You'll typically have multiple Feature nodes — one per major feature or user story.

Assets

Stores reference material. Supports image uploads, video uploads, PDF uploads, and Figma URL embeds. Uploaded assets can be annotated directly in the browser.

See Node types for the full field reference.

Edges / Connections

Edges are the lines drawn between nodes. They serve a functional purpose beyond visual organisation: connected nodes share context with each other's AI conversations.

When you chat on a Feature node that is connected to your Tech Stack node, the AI already knows your technology choices and tailors its questions accordingly.

See Connections for details.

Conversations

Each node has its own AI conversation. Conversations are persistent — they continue from where you left off across sessions.

The AI asks targeted questions based on the node type and uses the answers to extract structured data automatically. You never fill in a form — you just talk.

See AI conversations overview for details.

Structured Data

As you chat, the AI extracts key information and populates a structured data panel attached to each node. For example, chatting on an Audience node might auto-populate fields like primaryPersona, companySize, and painPoints.

Fields you edit manually are protected — the AI won't overwrite values you've set explicitly.

See Structured data for details.

PRDs

A PRD (Product Requirements Document) is the output Kommit generates from your canvas. It assembles all node data — conversations, structured fields, and optional repo context — into a comprehensive, structured document.

PRDs are versioned. Every generation is saved to history. You can regenerate as your spec evolves.

See Generating a PRD and Exporting for details.