All work

Project 09 / Obzidian

An AI app-building workspace with code, preview and deployment paths.

Project
Obzidian
Sector
AI App Builder / Developer Tools
Work type
Partner work
Capabilities
Web Development, SaaS Development, AI Agents, AI Automation
Status
AI app builder
Obzidian landing page with the prompt box that starts a new app

The project

An AI application builder with project dashboards, file trees, editors, preview services, chat-driven generation and deployment integrations.

The product context

A workspace for generating and shaping apps.

Obzidian is an AI app-building interface with project dashboards, editor routes, code preview surfaces and ideation paths.

Obzidian dashboard listing a demo workspace's projects with their status and last update
Obzidian dashboard listing a demo workspace's projects with their status and last update
The boundary

Generated code stays inspectable.

Generated files stay visible and editable next to the preview, so a person can review what the AI produced before it is deployed.

Obzidian editor with the project file tree, the code pane and an AI prompt bar
Obzidian editor with the project file tree, the code pane and an AI prompt bar
Builder surface

Generation, files and preview meet inside one workspace.

The product connects Clerk authentication, Supabase services, Sandpack and E2B preview paths, GitHub, Netlify and Stripe integrations, plus file, socket and SSE service modules.

The connected product
  1. 01Project dashboard
  2. 02Chat and ideation input
  3. 03File tree and editor
  4. 04Code preview services
  5. 05GitHub and deploy modules
  6. 06Billing and account routes
The entry point

One prompt starts the project.

The landing page leads with the prompt: describe the app, then continue into the workspace to shape it.

Obzidian landing page with the prompt box that starts a new app
Obzidian landing page with the prompt box that starts a new app
What changed

AI-assisted software creation in a controlled workspace.

Obzidian brings together developer-facing UX, generated code preview, service integrations and the operational concerns around letting AI create software in a controlled workspace.

What shipped

  • Prompt-based project creation
  • Project dashboard
  • File tree and code editor
  • Live code preview
  • GitHub and Netlify deployment
  • Stripe billing

Evidence

  • Dashboard, editor and landing screens from a local run with a demo workspace
  • AI generation, preview and deployment in one workspace

What changed

A user can describe an app, inspect and edit the generated files, preview them and push to deployment without leaving the workspace.

Next project

Maison Kayser Nigeria

Maison Kayser Nigeria homepage presenting the bakery and restaurant experience