Skip to main content
Kommit can import projects from a variety of sources. During import, Kommit analyzes your codebase to automatically detect your tech stack, database schema, features, and project structure — giving you a structured spec from day one.

Supported sources

SourceMethodWhat’s analyzed
GitHubConnect account or paste URLCode, tech stack, DB schema, features
GitLabPaste repository URLCode, tech stack, DB schema, features
BitbucketPaste repository URLCode, tech stack, DB schema, features
LovableConnect via GitHubCode, design, DB schema
BoltZip upload or GitHubCode, design, backend
Base44Connect via GitHubCode, design
VercelConnect via GitHubCode, design, backend, DB schema
FigmaPaste Figma file URLScreens, component structure
Zip fileUpload .zipCode, design, tech stack, DB schema

How it works

  1. Select a source — Choose where your project lives
  2. Connect or paste — Either connect your account (GitHub) or paste a URL
  3. Configure — Name your project and set visibility (public or private)
  4. Import — Kommit clones and analyzes your code in the background
Once complete, you’ll have a fully structured project with auto-detected nodes for tech stack, features, and database schema on your canvas.

Project visibility

When creating a project during import, you can choose:
  • Private (default) — Only workspace members can access the project
  • Public — Anyone with the link can view the project (useful for sharing specs with contractors or stakeholders)
You can change visibility later from project settings.

What Kommit analyzes

During import, Kommit examines your codebase to extract:
  • Tech stack — Detected from package.json, pyproject.toml, go.mod, Cargo.toml, Gemfile, composer.json, and other manifest files
  • Database schema — Parsed from Prisma schemas, Drizzle configs, and SQL migration files
  • Features — Inferred from route handlers, components, and module structure
  • Code structure — File organization, dependencies, and architecture patterns
All extracted data populates your canvas as structured nodes that you can refine with AI-guided conversations.