An AI-powered knowledge systems workspace — built as a portfolio demonstration of AI Knowledge Systems Design & Architecture capabilities.
Synaptica Knowledge Architecture is a suite of five AI-powered tools designed for documentation professionals, knowledge engineers, and technical writers who want to demonstrate or apply AI-driven thinking to knowledge management. Each tool reflects a core competency expected of an AI Knowledge Systems Designer/Architect.
Built as a portfolio project under the Synaptica Knowledge Systems brand, this workspace showcases practical, working implementations of AI applied to the full knowledge lifecycle — from creation and structuring to gap analysis, retrieval, and onboarding.
Intelligently searches across a pasted knowledge base using AI-ranked relevance scoring — no indexing or vector database required.
What it does:
- Accepts raw documentation or knowledge base content (Confluence exports, wikis, SOPs, policies)
- Accepts a natural language search query
- Returns ranked results with relevance scores, excerpts, and explanations of why each result matches
- Provides an AI-generated summary answering the query from the documents
Ideal for: Demonstrating AI-augmented knowledge retrieval, semantic understanding over keyword search, and explainable AI results.
Analyses existing documentation against real-world context (support tickets, user stories, or product descriptions) to identify what's missing — before users notice.
What it does:
- Accepts existing documentation and supplementary context (e.g. support ticket themes, user journey notes)
- Accepts an optional target audience specification
- Returns a coverage score (0–100%), a prioritised list of gaps (High / Medium / Low), suggested content for each gap, an overall summary, and strategic recommendations
- Helps prioritise documentation work by impact rather than guesswork
Ideal for: Demonstrating proactive knowledge strategy, AI-assisted content planning, and audit-grade thinking about documentation systems.
Transforms flat, dense documentation into structured, intent-based FAQ content — organised by topic and written the way users actually ask questions.
What it does:
- Accepts raw source documentation (manuals, specs, policies, guides)
- Accepts an optional audience specification and item count limit
- Returns categorised FAQ items with natural-language questions, concise answers, and keyword tags
- Presents results as an interactive, collapsible accordion by category
Ideal for: Demonstrating knowledge architecture thinking — structuring information around user intent rather than document structure.
A RAG-style (Retrieval-Augmented Generation) conversational AI that onboards new team members based on injected company documentation — your own knowledge base becomes the AI's only source of truth.
What it does:
- Creates persistent onboarding sessions with a trainee name, role, and a custom knowledge base (paste any wiki, policy set, or process document)
- Maintains a streaming chat interface with full conversation history
- The AI assistant answers questions using only the injected documentation, citing relevant sections where applicable
- Multiple sessions can be created and revisited
- Welcomes the trainee by name and role at session start
Ideal for: Demonstrating practical RAG implementation, conversational AI design, and the application of knowledge systems to real enterprise onboarding challenges.
A curated library of professional AI prompts built specifically for knowledge systems work, with a live sandbox to test, iterate, and save custom prompts.
What it does:
Library tab:
- Displays 8 built-in professional prompt templates across categories: Drafting, Summarization, Analysis, Quality, Onboarding, FAQ, Architecture
- Each prompt shows its template, description, category badge, and variables
- Any prompt can be loaded directly into the sandbox with one click
Sandbox / Test tab:
- Accepts any prompt template using
{{variable}}placeholder syntax - Auto-detects variables from the template and renders individual input fields
- Streams AI responses live as the model generates them
- Allows saving any tested prompt to your personal library
Built-in prompt categories:
| Prompt | Category | Purpose |
|---|---|---|
| Technical Procedure Writer | Drafting | Step-by-step procedures from raw notes |
| Audience-Specific Summarizer | Summarization | Technical content rewritten for specific audiences |
| Documentation Gap Prompter | Analysis | Identifies missing information in documents |
| Style Guide Checker | Quality | Flags content inconsistencies against style rules |
| Onboarding Script Generator | Onboarding | Onboarding guides for new team members |
| FAQ Question Generator | FAQ | Top questions and answers from documentation |
| Release Notes Writer | Drafting | User-friendly release notes from technical changes |
| Knowledge Base Structurer | Architecture | Taxonomy and article structure from raw content |
| Layer | Technology |
|---|---|
| Frontend | React 18, TypeScript, Vite |
| Styling | Tailwind CSS v4, Framer Motion |
| UI Components | shadcn/ui (Radix UI) |
| Routing | Wouter |
| State Management | TanStack React Query |
| Backend | Express 5, Node.js |
| Database | PostgreSQL + Drizzle ORM |
| AI / LLM | OpenAI GPT-5.2 (chat completions + streaming SSE) |
| API Contract | OpenAPI 3.1 + Orval codegen |
| Validation | Zod |
| Monorepo | pnpm workspaces |
Synaptica follows a contract-first full-stack architecture:
synaptica-knowledge-architecture/
├── artifacts/
│ ├── api-server/ # Express 5 REST API
│ │ └── src/routes/
│ │ └── knowledge/ # All 5 AI tool endpoints
│ └── knowledge-workspace/ # React + Vite frontend
├── lib/
│ ├── api-spec/ # OpenAPI 3.1 specification
│ ├── api-client-react/ # Generated React Query hooks
│ ├── api-zod/ # Generated Zod validation schemas
│ ├── db/ # Drizzle ORM schema + DB connection
│ └── integrations-openai-ai-server/ # OpenAI integration layer
All API contracts are defined first in OpenAPI, then generated into type-safe React Query hooks and Zod schemas — ensuring the frontend and backend never drift apart.
All endpoints are under /api/knowledge/:
| Method | Endpoint | Tool |
|---|---|---|
POST |
/knowledge/search |
Knowledge Search |
POST |
/knowledge/analyze-gaps |
Gap Analyzer |
POST |
/knowledge/build-faq |
Smart FAQ Builder |
GET |
/knowledge/onboarding/sessions |
List Onboarding Sessions |
POST |
/knowledge/onboarding/sessions |
Create Onboarding Session |
GET |
/knowledge/onboarding/sessions/:id/messages |
Get Chat History |
POST |
/knowledge/onboarding/sessions/:id/messages |
Send Message (SSE) |
GET |
/knowledge/prompts |
List Prompt Templates |
POST |
/knowledge/prompts |
Save Custom Prompt |
DELETE |
/knowledge/prompts/:id |
Delete Prompt |
POST |
/knowledge/prompts/test |
Test Prompt (SSE) |
- Node.js 20+
- pnpm 9+
- PostgreSQL database (or Replit's built-in DB)
- OpenAI API key (or Replit AI Integrations access)
# Clone the repository
git clone https://github.com/Miru-Maria/synaptica-knowledge-architecture.git
cd synaptica-knowledge-architecture
# Install dependencies
pnpm install
# Set environment variables
cp .env.example .env
# Add DATABASE_URL, AI_INTEGRATIONS_OPENAI_BASE_URL, AI_INTEGRATIONS_OPENAI_API_KEY
# Push database schema
pnpm --filter @workspace/db run push
# Start the API server
pnpm --filter @workspace/api-server run dev
# Start the frontend (in another terminal)
pnpm --filter @workspace/knowledge-workspace run devThis project was built as part of the Synaptica Knowledge Systems portfolio — a personal brand and freelance offering focused on the intersection of AI, technical writing, and knowledge architecture.
It demonstrates practical, deployable AI systems built around the core competencies of an AI Knowledge Systems Designer/Architect:
- Structuring and retrieving knowledge semantically
- Identifying and filling documentation gaps strategically
- Designing AI-augmented onboarding experiences
- Engineering reusable prompt systems for knowledge work
- Building full-stack AI tools with production-quality architecture
MIT — feel free to explore, fork, and build on this.
Synaptica Knowledge Systems — Designing the intelligence layer of the modern organisation.