AI-powered log diagnostics with frictionless access and controlled usage
Backend: FastAPI | Auth & DB: Supabase | AI: Google Gemini | Hosted on Render
LogAI is a production-deployed AI-powered log analysis platform designed to help developers, DevOps engineers, and SRE teams diagnose system issues efficiently.
It converts raw log files into structured, actionable diagnostic reports using large language models, eliminating the need to manually inspect large volumes of logs.
Version 2.1 introduces anonymous access with controlled usage, improving onboarding while maintaining system cost protection and stability.
Modern systems generate large volumes of logs that are difficult to analyze under time constraints.
LogAI enables:
- Faster incident diagnosis
- Reduced debugging effort
- Clear root cause identification
- Actionable remediation guidance
- Controlled system usage with cost protection
- Seamless transition from anonymous to authenticated usage
The platform functions as an intelligent diagnostic layer for production environments.
- Identifies critical failures and anomalies
- Provides structured explanations
- Suggests actionable remediation steps
- Detects recurring or suspicious patterns
- Server-Sent Events (SSE) for live updates
- Chunk-level analysis visibility
- Progressive result delivery
- No login required for initial usage
- Daily usage limits enforced per user/session
- Backend-controlled usage validation before AI execution
- Seamless transition to authenticated usage after limits
- Supabase Auth integration
- Server-side ES256 JWT verification
- JWKS-based key validation
- Strict per-user data isolation
- Available for authenticated users
- Timestamped reports with structured summaries
- Historical retrieval and review
- Severity classification (critical, error, warning, info)
- Context-preserving chunking
- Cost-aware AI orchestration
- Efficient token usage
User (Anonymous or Authenticated)
│
▼
Frontend (Netlify)
│
├── Anonymous Session (anon_id)
├── Optional Authentication (Supabase)
├── File Upload + SSE Streaming
│
▼
FastAPI Backend (Render)
│
├── Usage Guard Layer (v2.1)
│ ├── Anonymous usage tracking
│ └── Authenticated usage tracking
│
├── JWT Verification (JWKS)
├── Log Preprocessing Engine
├── Chunk Management
├── AI Orchestration (Gemini Flash)
└── Secure Storage (Supabase)
│
▼
Google Gemini Flash
- Additive system design – non-destructive feature evolution
- Zero trust on client-side logic
- Backend-enforced usage and security controls
- Stateless AI processing with controlled state layers
- Service-role database operations – never exposed to client
- Cloud-native deployment architecture
| Layer | Implementation |
|---|---|
| Token Verification | ES256 JWT with JWKS key resolution |
| Usage Enforcement | Backend-enforced daily limits |
| Database Access | Supabase service-role (never client-exposed) |
| API Protection | CORS-restricted access |
| Configuration | Environment-based secret management |
LogAI v2.1 maintains production-grade security while enabling flexible user access.
| Layer | Technologies |
|---|---|
| Backend | FastAPI · Python · LangChain · Google Gemini Flash · Supabase · Uvicorn |
| Frontend | Vanilla JavaScript · Marked.js · Highlight.js · Responsive CSS |
| Infrastructure | Netlify (Frontend) · Render (Backend) · Supabase (Auth & DB) |
- Application logs – Python, Node.js, Java
- System logs – Linux, systemd, kernel
- Web server logs – Nginx, Apache
- Container logs – Docker, Kubernetes
- Database logs
- Free-tier refinement and usage optimization
- Improved UX and feedback mechanisms
- Export capabilities (PDF / structured reports)
- Enhanced error handling and observability
- Billing and subscription system
- Team and organization support
- API access for external integrations
- Usage analytics dashboard
- Background job processing
| Version | Description |
|---|---|
| v1.0 | Streaming AI Log Analyzer |
| v2.0 | Authenticated Multi-User SaaS Architecture |
| v2.1 | Anonymous Access with Usage Control and Improved UX (Current) |
Harsh Pathak GitHub
If you find LogAI useful, consider starring the repository.