RAG assistants
A chatbot that answers only from your data and cites the source. Just like Robin - it does not make things up.
jonas@pilsen:~/build$ ./build_ai_app.sh
RAG assistants over your documents, agents, internal tools, or AI features inside an existing product. From idea to a running product, not to a slide deck. I build it myself, so I know what is realistic.
What I can build
A chatbot that answers only from your data and cites the source. Just like Robin - it does not make things up.
Automations that act on their own - briefings, content, replies, hooked into your tools.
Dashboards and apps built around your process, not around a template. For where a spreadsheet stops being enough.
I add a sensible AI feature to the app you already have - search, summarising, classification.
How it goes
What you are dealing with and what the app actually has to do. No filler.
A quick working core you can actually poke at, not a mockup.
We iterate on your real data until it fits.
I put it into production and hand it over.
Stack
Proof, not promises
An internal AI assistant for companies. RAG over company documents, roles, source citations. Built entirely from scratch.
Breaking down viral Reels with Gemini plus concept generation with Claude. Next.js, SSE pipeline.
Pricing