Reverse engineered prompt

I want this turned into a local backend for AI apps that feels like a Supabase style service, but for knowledge instead of regular app data. The main thing I need is a working setup where I can start it with Docker and yarn, connect an OpenAI key, create a project, add a data source, and have it ingest and index content so an app can use it for search, memory, chat, or RAG style features.

Please wire up whatever is needed so the repo actually runs cleanly on a current Node LTS setup, with sensible env setup from the example file and clear commands for starting everything. If the indexing flow depends on the Temporal dashboard, make sure that works too and is easy to check locally.

This repo is archived, so if anything is outdated, use the existing code and README as the base and look up current docs online only if you need to unblock the setup. I would also love a tiny end to end example that proves ingestion and querying are working.

Want more depth? Deep Reverse