AI assistants & workflow automation · Engineering case study
Lamp: Lamp Technology Stack
Application-scoped AI orchestration, durable conversation storage and separate workspace integrations.
Inside Lamp.
Explore LampThe Colyf integration uses a TypeScript application harness with server-side model APIs, authenticated action routes and PostgreSQL conversation storage.
React and React Native provide the web and mobile chat experiences. Supported document attachments, follow-up questions and review cards connect the conversation to the application workflow.
The separate workspace agent for Telegram, WhatsApp and Slack uses a Python-based Hermes runtime and narrowly configured integrations. Its memories and tool access stay separate from Colyf's application context.
Model providers and deployment settings can change without turning provider names or model output into authority over application permissions. Credentials remain outside the public frontend.
- TypeScript application harness and server-side model APIs
- PostgreSQL for private saved conversations
- React / React Native clients and a separate Python workspace integration