Internal Tools Workspace Framework Specification
December 24, 2025 ยท View on GitHub
Problem Statement
Multiple standalone AI-powered utilities have been (or will be) developed independently. These tools:
- Are too small to warrant individual apps/websites
- Share common needs: auth, API key management, cross-platform access
- Currently lack a unified deployment and access strategy
What's needed: A framework/shell that bundles these existing utilities into a cohesive authenticated workspace.
What This Spec Covers
This is a specification for the wrapper, not the tools inside it:
- Deployment framework
- Authentication layer
- Navigation/routing shell
- Shared configuration management
- Module integration pattern
Core Requirements
1. Authentication Layer
- Single login protects all modules
- Prevents unauthorized API credit usage
- Simple to implement (not enterprise-grade)
- Options: OAuth (Google), basic auth, session-based
2. Module Shell / Navigation
- Sidebar or menu-based navigation
- Desktop: persistent sidebar
- Mobile (Android): hamburger menu or bottom nav
- Modules render within the shell's content area
- Smooth transitions between modules
3. Shared Environment Variables
Centralized configuration for:
GEMINI_API_KEYOPENROUTER_API_KEYREPLICATE_API_KEY- Google API credentials
- Other service keys as needed
Modules inherit these without individual configuration.
4. Cross-Platform Support
- Desktop: Browser-based
- Mobile: Android phone via browser (responsive design)
- Optional: PWA for app-like mobile experience
5. Module Integration Pattern
How existing/new tools plug into the workspace:
- Framework-agnostic (React, Vue, vanilla JS modules should all work)
- Self-contained module directories
- Standard interface for mounting into the shell
- Module isolation (one module's failure doesn't crash others)
6. Deployment
Available infrastructure:
- VPS: Self-hosted option
- Vercel: Managed platform with env var sharing
- Home server: Docker-based deployment
- N8N: Backend workflow orchestration (some modules may call N8N endpoints)
Non-Requirements
- Not building the AI utilities themselves (already handled)
- Not a no-code platform (Retool, Appsmith, etc.)
- Not a conversational AI interface
- Not a SaaS for external users
- Not tied to a specific AI vendor's ecosystem
Success Criteria
- Integration speed: New module integrates in <1 hour
- Single auth: One login, all modules accessible
- Cross-platform: Works on desktop browser and Android
- Shared config: Environment variables accessible to all modules without per-module setup
- Framework flexibility: Can drop in React, Vue, or other frontend code
- Minimal overhead: The wrapper adds value, not complexity
Evaluation Criteria for Framework Selection
- How easily can custom modules be added?
- Does it provide auth out of the box or require building it?
- How does it handle shared environment variables?
- Mobile/responsive support?
- Deployment options (Vercel, Docker, VPS)?
- Lock-in risk - can modules be extracted if framework is abandoned?
- Learning curve for initial setup?