vectorflow

Why we included this project

VectorFlow is for teams that need to move large volumes of documents into a vector database without building the ingestion plumbing themselves. You submit raw files through a simple API and it handles chunking and embedding, then writes the vectors to a store you choose, with Pinecone, Qdrant, and Weaviate supported out of the box. Under the hood it runs on a job queue backed by Postgres, RabbitMQ, and MinIO, so heavy batches get processed reliably instead of failing partway through. If you are putting together RAG-backed search or retrieval, it works as a self-hosted ingestion layer you can run with docker compose locally or scale on Kubernetes. Keep in mind it is still an MVP, so check that the current feature set covers your file types and target vector database before relying on it in production.