embedbase

Why we included this project

Embedbase wraps retrieval-augmented plumbing in a small client: add documents, search them semantically, and generate a text answer with a few methods. That gets a proof of concept against a vector store and an LLM running in minutes, without hand-wiring the usual stack, like pgvector, Supabase, and the OpenAI SDK. Because this repository holds the code behind the hosted service, it also works as a reference if you'd rather run the integration layer yourself than call the cloud API. The appeal is one code path for semantic search and text generation, instead of several services that each carry their own setup.