AIclipse

AIclipse is a polyglot microservices platform for AI image detection. It combines browser-facing BFFs, a gateway trust boundary, AI inference, media storage, billing, community features, model lifecycle management, and an email worker into one Kubernetes-based system.

What It Does

At a high level, AIclipse gives you:

• AI image detection with verdict, confidence, and model versioning
• image history and private/public media management
• a community feed with posts, comments, votes, moderation, and notifications
• subscription and billing flows
• model upload, deployment, training, and confidence scoring workflows

The current runtime is built around nine services:

Service	Stack	Role
auth	FastAPI + MongoDB	identity, JWTs, API keys, admin actions
gateway	FastAPI	trust boundary, auth validation, proxying
client	Flask	main browser BFF and server-rendered site
community	Next.js	/community UI, internal routes, adminBFF
billing	FastAPI + MongoDB + Stripe	checkout and billing state
media	FastAPI + MongoDB + MinIO/S3	image metadata and presigned URLs
detector	FastAPI + PyTorch/Transformers	AI inference service
model-cycle	ASP.NET Core + SQLite + Python	model lifecycle and training
email-worker	FastAPI + Redis Streams + Gmail	email dispatch worker

Screenshots

Landing Page

Detection Flow

Profile

Community Feed

Admin Management

_

Architecture

Public development routing:

• / -> client
• /community -> community
• /api/* -> gateway with ingress rewrite
• storage.aiclipse.local -> MinIO/object storage

Core storage:

• MongoDB for documents
• Redis for streams, debounce buffers, and counters
• MinIO for S3-compatible object storage
• SQLite inside model-cycle for model metadata

Primary architecture contract:

• ARCHITECTURE.md

Repository Layout

Top-level directories:

• auth/
• billing/
• client/
• community/
• detector/
• email-worker/
• gateway/
• infra/
• media/
• model-cycle/
• swagger-docs/

Key implementation areas:

• client/routes, client/services, client/templates, client/static
• community/app, community/lib, community/scripts
• gateway/app/core, gateway/app/routers
• auth/app/core, auth/app/db, auth/app/routers, auth/app/services
• detector/detector_modules, detector/routers
• model-cycle/ModelCycle/Controllers, Services, Repositories
• infra/k8s, infra/k8s-dev, infra/k8s-prod

Getting Started

For the full operational guide, local setup steps, exact run commands, and test/build details, read:

• Readme.txt

Fast path:

1. Enable Kubernetes in Docker Desktop.
2. Add 127.0.0.1 aiclipse.local storage.aiclipse.local to your hosts file.
3. Install ingress-nginx.
4. Run skaffold dev from the repository root.
5. Open http://aiclipse.local.

Build And Test

There is no root Makefile. The intended development path is repository-root skaffold dev.

Common local test entry points:

• Python services: python -m pytest tests -q
• client: npm test
• community: npm test
• model-cycle: dotnet test Tests/Tests.csproj --no-restore --verbosity minimal

CI/CD entry points live in:

• .github/workflows/

Contracts And Source Of Truth

When in doubt, use this order:

1. ARCHITECTURE.md
2. source code and Kubernetes manifests
3. OpenAPI contracts in swagger-docs/

Useful files:

• ARCHITECTURE.md
• Readme.txt
• skaffold.yaml
• infra/kustomization.yaml
• swagger-docs/

Final Guidance

For normal development, prefer:

• running the full stack with skaffold dev
• treating ARCHITECTURE.md as the current contract
• using Readme.txt as the operational runbook
• using Kubernetes manifests as the deployment truth