Documentation
Docs
Everything you need to generate ideas, use your workspace, and configure the app.
Getting started
- Sign in with GitHub or Google.
- Click Create Idea to open the console and fill in your profile or context.
- Submit the form — you'll land on the generation page with live logs.
- When complete, explore your workspace: ranked projects, attachments, and per-project tabs.
Past sessions are available from your Dashboard.
Idea intake
The console form supports three input modes:
- Developer profile — skills, GitHub username, time available, career goals
- Freeform context — hackathon themes, constraints, niche spaces
- Both — combine profile and context
Seed idea: optional textarea — all generated ideas extend or relate to your concept.
File uploads: images, PDF, txt, md, json, csv (max 5 files, 5 MB each). Uploaded files are stored and viewable from the workspace.
Without seed or files → 3 ideas. With seed and/or files → 5 ideas (closest match, 3 variations, 1 adjacent).
Toggles:
- Agentic coding — skills become optional when building with AI agents
- AI in the product — allow LLM/agent features inside the product itself
Workspace
Each idea generation session creates a workspace containing ranked projects. From the workspace you can:
- View generation status and summary callouts
- Browse and open individual project tabs
- Generate another idea within the same session
- View and download uploaded attachments
The Dashboard lists all your past sessions with status badges and quick links.
Project tabs
Each project in a workspace has five tabs:
- Overview — full idea card with pitch, stack, MVP scope, roadmap, and monetization
- Agent — auto-generated AGENTS.md; copy, download, or regenerate
- Canvas — React Flow architecture diagram; editable nodes/edges with real-time validation and design system panel
- MVP — live UI preview powered by v0 (requires Agent tab to be ready)
- Chat — streaming per-project chat with full project context
Environment variables
| Variable | Required | Description |
|---|---|---|
| LLM_API_KEY | Yes* | API key for your LLM provider |
| LLM_BASE_URL | No | Custom base URL (OpenAI, Groq, Ollama, etc.) |
| LLM_MODEL | No | Model name (default: gpt-4o-mini) |
| NEXT_PUBLIC_SUPABASE_URL | Yes | Supabase project URL |
| NEXT_PUBLIC_SUPABASE_ANON_KEY | Yes | Supabase anon key |
| DATABASE_URL | Yes | Postgres connection (Supabase pooler) |
| GITHUB_TOKEN | No | Optional — enriches GitHub profile fetch during generation |
| V0_API_KEY | No | Required for MVP preview generation |
*Or use OPENAI_API_KEY as a fallback alias for LLM_API_KEY.
Limits
- 5 workspace creations per hour per IP
- Max 5 file uploads, 5 MB each
- MVP preview requires Agent instructions to be ready and a valid
V0_API_KEY - Image extraction uses your configured LLM — vision-capable models (e.g. gpt-4o-mini) work best