What Whilst Does
Knowledge Extraction
Every Slack conversation is automatically analyzed and turned into structured documentation with topics, decisions, and action items.
Semantic Search
Find documents by meaning, not just keywords, using pgvector embeddings and hybrid search with Reciprocal Rank Fusion.
AI Sidekick
Chat with AI about your docs, get intelligent suggestions, and automate document tasks from the web app.
Relationship Intelligence
Tracks who works with whom on what topics, building a collaboration graph that makes future responses smarter.
Key Capabilities
- Automatic Knowledge Extraction — Mention
@whilstin Slack and the pipeline captures, analyzes, and documents the thread - Hybrid Search (RRF) — Combines semantic similarity and full-text search for best-in-class accuracy
- AI-Powered Analysis — Extracts topics, decisions, action items, and sentiment from threads
- Context-Aware Responses — Bot understands your projects, collaborators, and conversation history
- MCP Integrations — Query GitHub, Linear, and Notion through natural language
- MCP Server — Enable AI coding assistants (Cursor, Claude Desktop, VS Code) to query and manage your docs
- Real-time Collaboration — WebSocket sync across clients via AWS AppSync Events
Architecture at a Glance
Tech Stack
| Layer | Technology |
|---|---|
| Frontend | Next.js 14 (App Router), React, Tailwind, shadcn/ui, TipTap |
| Backend | Node.js, TypeScript, AWS Lambda |
| Database | PostgreSQL (Neon) with pgvector |
| AI | OpenAI GPT-4o, text-embedding-3-small |
| Auth | WorkOS (SSO, email verification) |
| Real-time | AWS AppSync Events (WebSocket) |
| Infrastructure | AWS CDK, Vercel, GitHub Actions |

