Node Types
The six node types in Kommit and the structured data each one captures.
Node Types
Kommit has six node types. Each represents a distinct dimension of your product specification and holds its own AI conversation. The structured data extracted from those conversations feeds directly into your PRD.
Project
The root node for your product. Every project starts with one Project node and you can only have one.
Captures:
- Product name
- Short description
- Problem statement
- High-level goals
- Target launch timeline
Role in the PRD: Becomes the executive summary and sets the framing for every other section.
Audience
Describes who you are building for. You can have multiple Audience nodes to represent distinct user segments.
Captures:
- Primary persona (name, role, background)
- Company size / segment
- Key pain points
- Jobs to be done
- Success metrics (how does this user know the product worked?)
Role in the PRD: Populates the Target Users section and informs feature prioritisation.
Competitors
Maps the competitive landscape. Add one node per competitor, or use a single node to cover multiple.
Captures:
- Competitor names
- Their core strengths
- Their notable weaknesses
- Gaps your product addresses
- Positioning differentiation
Role in the PRD: Populates the Competitive Analysis section.
Tech Stack
Records the planned (or existing) technical architecture. One Tech Stack node per project is typical.
Captures:
- Frontend framework and libraries
- Backend runtime and framework
- Database(s)
- Infrastructure and hosting
- Third-party services and APIs
- Authentication approach
Role in the PRD: Populates the Technical Architecture section. Also used as context in Feature node conversations so the AI can tailor acceptance criteria and complexity estimates to your actual stack.
Features
Describes a single product capability. Add one Feature node per major feature or user story.
Captures:
- Feature name
- User-facing description
- User story (
As a [role], I want to [action] so that [outcome]) - Priority (P0 = must-have, P1 = should-have, P2 = nice-to-have, P3 = future)
- Acceptance criteria (bulleted list)
- Estimated complexity (low / medium / high)
- Dependencies on other features
Role in the PRD: Each Feature node becomes a section in the Features & Requirements chapter of the PRD.
Connect Feature nodes to your Tech Stack and Audience nodes. The AI uses that context to write more precise acceptance criteria and flag technical constraints you may not have considered.
Assets
Holds reference material — designs, screenshots, recordings, and specs. Attach an Assets node to any Feature or to the project as a whole.
Supports:
- Image uploads (PNG, JPG, GIF, WebP)
- Video uploads (MP4, WebM)
- PDF uploads
- Figma URL embeds (iframe preview)
Additional capabilities:
- Folders — organise files within the node
- Annotations — draw arrows, rectangles, and text directly on images in-browser
- Versions — upload a new version of any file and revert to previous versions
- Descriptions — add a text description to each asset; these are embedded into the knowledge base and used as context in AI conversations
Role in the PRD: Asset descriptions and annotations are surfaced in the PRD's Design & Assets section. Images are also sent to the AI as visual context during Feature conversations.
Node editing
Click a node to open its side panel. You can edit the node's title and description directly in the panel header. Structural fields (like priority or tech stack choices) are populated through conversation and can be viewed in the Extracted Data panel.
All edits auto-save with the canvas.