finds.dev← search

// the find

weaviate/Verba

★ 7,714 · Python · BSD-3-Clause · updated Jun 2026

Retrieval Augmented Generation (RAG) chatbot powered by Weaviate

Verba is a RAG chatbot UI built on Weaviate, letting you ingest documents and query them via an LLM of your choice. It's for developers or teams who want a self-hosted, no-code-required RAG interface without building one from scratch. The project is now archived and officially discontinued.

The provider matrix is genuinely broad — embeddings and generation from Ollama, OpenAI, Cohere, Anthropic, VoyageAI, Groq, and others are all first-class, not afterthoughts. Chunking options are solid: token, sentence, semantic, recursive, and format-aware chunkers for HTML, Markdown, JSON, and code. Docker + docker-compose setup is straightforward and works without fighting the tool. The 3D vector visualizer is a nice debugging aid for understanding what your embeddings are actually doing.

The project is archived and dead — no security patches, no dependency updates, no bug fixes. Weaviate Embedded doesn't run on Windows, so a significant chunk of potential users are forced into Docker or cloud deployment immediately. Single-user only by design, with no plans for multi-user support even before discontinuation, which rules it out for any team use case. Reranking and RAG evaluation — the features that would actually make this useful for production quality tuning — were listed as 'planned' and never shipped.

View on GitHub →

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →