Knowledge Hub for Teams

Knowledge Hub is an internal Q&A and knowledge-base platform for teams, inspired by Stack Overflow and Laracasts. It combines a full Q&A workflow (questions, answers, voting, reputation, categories, tags, comments, bookmarks, notifications) with real-time updates, optional AI-generated answers, and an AI-powered knowledge base (RAG) over project documents and emails. All documentation is in English.

What This Project Is

• Q&A core — Members ask questions, post answers, vote, accept answers, and earn reputation. Content supports Markdown and image uploads.
• Taxonomy — Admins manage categories (hierarchical) and tags; members filter and full-text search questions.
• Social — Comments on questions/answers, bookmarks, and in-app notifications when someone answers your question.
• Real-time — New answers, votes, and notifications appear live via Laravel Reverb (WebSockets).
• AI (optional) — Provider-agnostic LLM layer: generate draft answers on questions; all calls audited. Optional auto-answer on new questions.
• Knowledge Base (Phase I) — Projects group documents (PDF/DOCX/TXT) and emails. Upload or add manually; vector embeddings and RAG let you “Ask AI” with contextual answers and citations. Export to Markdown or PDF; activity log for uploads, processing, RAG, and exports.

Stack: Laravel 12, Inertia + Vue 3, Tailwind, Docker, PostgreSQL with pgvector. Phases A–I are implemented (see below).

Documentation (/docs)

Each phase is documented in a single file (spec, user test plan, and dev notes merged). Reference doc for migrations and assumptions:

Phase	File	Content
A	phase-a-initial-setup.md	Foundation: Laravel 12, Inertia + Vue 3 + Tailwind, Docker, Pint, test skeleton, README
B	phase-b-auth-rbac.md	Auth, RBAC (Admin/Moderator/Member), admin skeleton, user management
C	phase-c-qa-core.md	Q&A CRUD, Markdown, image uploads
D	phase-d-voting-accepted-reputation.md	Voting, accepted answer, reputation
E	phase-e-categories-tags-filters-search.md	Categories, tags, filters, full-text search
F	phase-f-comments-bookmarks-notifications.md	Comments, bookmarks, in-app notifications
G	phase-g-realtime-reverb.md	Real-time (Reverb): answers, votes, notifications
H	phase-h-ai-integration.md	AI layer (OpenAI/Anthropic/Gemini), Generate AI Answer, audit
I	phase-i-rag-knowledge-base.md	Projects, documents/emails, RAG, export, activity log
—	migrations-and-seeding.md	Migration order, seeding, known assumptions
—	troubleshooting.md	Common fixes and recovery steps

Quick Start (First Time)

1. Prerequisites: Docker (with Compose), Git. Ensure ports 8080, 5173, and 5432 are free.
2. Configure: Copy backend/.env.example to backend/.env. Set APP_URL=http://localhost:8080 (and optionally AI/Reverb vars; see Environment variables).
3. Run: Build and start services, then bootstrap the app.
4. Optional: Start a queue worker for Phase I document processing and Reverb for real-time.
5. Open: http://localhost:8080 — log in with seeded users (e.g. admin@knowledge-hub.test / password).

Commands (bash)

cp backend/.env.example backend/.env
docker compose up -d --build
docker compose exec app composer install
docker compose exec app php artisan key:generate
docker compose exec app php artisan migrate
docker compose exec app php artisan db:seed
docker compose exec app php artisan storage:link

Commands (PowerShell)

Copy-Item backend/.env.example backend/.env
docker compose up -d --build
docker compose exec app composer install
docker compose exec app php artisan key:generate
docker compose exec app php artisan migrate
docker compose exec app php artisan db:seed
docker compose exec app php artisan storage:link

Optional workers (run in separate terminals):

docker compose exec app php artisan queue:work
docker compose exec app php artisan reverb:start

Requirements

• Docker Desktop (or Docker Engine) with Compose
• Git

Running with Docker

• App (Nginx): http://localhost:8080
• Vite dev server: http://localhost:5173
• PostgreSQL: localhost:5432

Common commands:

docker compose logs -f --tail=100
docker compose exec app php artisan route:list
docker compose exec app php artisan test
docker compose exec node npm run build
docker compose down

Environment variables

Edit backend/.env as needed. Key variables:

• App: APP_URL (e.g. http://localhost:8080)
• Database: DB_* — DB_PASSWORD must match POSTGRES_PASSWORD in docker-compose.yml
• Queue: QUEUE_CONNECTION — use database (default) or redis; Phase I document processing runs via queue jobs
• Reverb (Phase G): REVERB_* for real-time broadcasting
• AI (Phase H / Phase I): AI_PROVIDER (openai, anthropic, gemini, or mock), AI_ENABLED, OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY; for RAG embeddings add AI_EMBEDDING_MODEL and AI_EMBEDDING_DIMENSION (e.g. 1536 for OpenAI text-embedding-3-small). See docs/phase-h-ai-integration.md and docs/phase-i-rag-knowledge-base.md.

All AI calls (including RAG embeddings and chat) are audited in ai_audit_logs.

Database

Migrations run in filename order. Phase I adds projects, knowledge_items, knowledge_chunks (with pgvector), rag_queries, and activity_logs. See docs/migrations-and-seeding.md for full order and reseed instructions.

docker compose exec app php artisan migrate
docker compose exec app php artisan db:seed

Troubleshooting

For common fixes and recovery steps, see docs/troubleshooting.md.

Architectural decisions

• PostgreSQL + pgvector: used for vector embeddings and similarity search in Phase I RAG.
• Inertia + Vue 3: SPA-like UI with Laravel routing and server-side conventions.
• AI layer (Phase H) is provider-agnostic; all AI usage (including RAG) is logged in ai_audit_logs.