Lena — AI Orchestration Workspace
In-development AI workspace connecting goals, private knowledge and on-demand analysis, with persistent results and human review.
The challenge
AI-assisted work needs more than isolated conversations: objectives, supporting documents, results and human decisions need to stay connected. I am building Lena to organize that workflow while keeping the owner in control.
The solution
I built a modular TypeScript and Node.js platform with Supabase authentication and PostgreSQL persistence. It connects goals and success criteria to private documents, lexical search with source citations, user-requested research and analysis jobs, chat history and a progress dashboard. Model adapters check estimated text-generation costs before execution, while stored results and job events support review and follow-up.
The outcome
An assisted version with integrated goals, documents, internal jobs and dashboard, developed and validated locally. The current release produces analysis and proposals; it does not independently execute code, publish content or perform financial actions. Development is ongoing, the source repository remains private and no public web demo is available yet. The cover is an illustration, not a screenshot of the application.