Case study
KairOS
AI agents that deliver a daily market briefing
A multi-agent system built for a client that collects launches and discussions from Reddit, Hacker News and Product Hunt and delivers a morning report.

01Overview
A considered experience.
Built for real people.
The client wanted to start every day knowing what had happened in their market without reading a dozen sites. KairOS runs a team of agents overnight: they collect new launches and discussions from Reddit, Hacker News, Product Hunt and Indie Hackers, store them in vector memory, link them in a knowledge graph and score each opportunity on demand, competition and effort. Each morning a report agent writes a daily briefing with the top opportunity, market discoveries, competitor launches and keyword ideas, and sends it to Slack or Discord, with a dashboard to explore everything behind it.
Built for a client as a freelance project.
02Key features & capabilities
Vector memory and a knowledge graph of everything collected
Opportunity scoring on demand, competition and effort
Daily morning report delivered to Slack or Discord
Dashboard with live agent logs
03Tech stack
04Technical architecture
The technology behind the experience.
- Separate agents for research, memory, decisions, reports and notifications
- FastAPI service with async SQLAlchemy and a Qdrant vector store
- Scheduled daily run with webhook delivery and a local fallback
- Docker Compose for the full stack
05Explore the product

Let?s work together
Interested in working
together?
I?m open to new projects and collaborations.
Let?s build something great.